Configuring GFS2 file systems
Planning, administering, troubleshooting, and configuring GFS2 file systems in a high availability cluster
Abstract
Providing feedback on Red Hat documentation
We appreciate your feedback on our documentation. Let us know how we can improve it.
Submitting feedback through Jira (account required)
- Log in to the Jira website.
- Click Create in the top navigation bar
- Enter a descriptive title in the Summary field.
- Enter your suggestion for improvement in the Description field. Include links to the relevant parts of the documentation.
- Click Create at the bottom of the dialogue.
Chapter 1. Planning a GFS2 file system deployment
The Red Hat Global File System 2 (GFS2) file system is a 64-bit symmetric cluster file system which provides a shared name space and manages coherency between multiple nodes sharing a common block device. A GFS2 file system is intended to provide a feature set which is as close as possible to a local file system, while at the same time enforcing full cluster coherency between nodes. To achieve this, the nodes employ a cluster-wide locking scheme for file system resources. This locking scheme uses communication protocols such as TCP/IP to exchange locking information.
In a few cases, the Linux file system API does not allow the clustered nature of GFS2 to be totally transparent; for example, programs using POSIX locks in GFS2 should avoid using the GETLK
function since, in a clustered environment, the process ID may be for a different node in the cluster. In most cases however, the functionality of a GFS2 file system is identical to that of a local file system.
The Red Hat Enterprise Linux (RHEL) Resilient Storage Add-On provides GFS2, and it depends on the RHEL High Availability Add-On to provide the cluster management required by GFS2.
The gfs2.ko
kernel module implements the GFS2 file system and is loaded on GFS2 cluster nodes.
To get the best performance from GFS2, it is important to take into account the performance considerations which stem from the underlying design. Just like a local file system, GFS2 relies on the page cache in order to improve performance by local caching of frequently used data. In order to maintain coherency across the nodes in the cluster, cache control is provided by the glock state machine.
Make sure that your deployment of the Red Hat High Availability Add-On meets your needs and can be supported. Consult with an authorized Red Hat representative to verify your configuration prior to deployment.
1.1. GFS2 file system format version 1802
As of Red Hat Enterprise Linux 9, GFS2 file systems are created with format version 1802.
Format version 1802 enables the following features:
-
Extended attributes in the
trusted
namespace ("trusted.* xattrs") are recognized bygfs2
andgfs2-utils
. -
The
rgrplvb
option is active by default. This allowsgfs2
to attach updated resource group data to DLM lock requests, so the node acquiring the lock does not need to update the resource group information from disk. This improves performance in some cases.
Filesystems created with the new format version will not be able to be mounted under earlier RHEL versions and older versions of the fsck.gfs2
utility will not be able to check them.
Users can create a file system with the older format version by running the mkfs.gfs2
command with the option -o format=1801
.
Users can upgrade the format version of an older file system running tunegfs2 -r 1802 device
on an unmounted file system. Downgrading the format version is not supported.
1.2. Key GFS2 parameters to determine
There are a number of key GFS2 parameters you should plan for before you install and configure a GFS2 file system.
- GFS2 nodes
- Determine which nodes in the cluster will mount the GFS2 file systems.
- Number of file systems
- Determine how many GFS2 file systems to create initially. More file systems can be added later.
- File system name
-
Each GFS2 file system should have a unique name. This name is usually the same as the LVM logical volume name and is used as the DLM lock table name when a GFS2 file system is mounted. For example, this guide uses file system names
mydata1
andmydata2
in some example procedures. - Journals
-
Determine the number of journals for your GFS2 file systems. GFS2 requires one journal for each node in the cluster that needs to mount the file system. For example, if you have a 16-node cluster but need to mount only the file system from two nodes, you need only two journals. GFS2 allows you to add journals dynamically at a later point with the
gfs2_jadd
utility as additional servers mount a file system. - Storage devices and partitions
-
Determine the storage devices and partitions to be used for creating logical volumes (using
lvmlockd
) in the file systems. - Time protocol
Make sure that the clocks on the GFS2 nodes are synchronized. It is recommended that you use the Precision Time Protocol (PTP) or, if necessary for your configuration, the Network Time Protocol (NTP) software provided with your Red Hat Enterprise Linux distribution.
The system clocks in GFS2 nodes must be within a few minutes of each other to prevent unnecessary inode time stamp updating. Unnecessary inode time stamp updating severely impacts cluster performance.
You may see performance problems with GFS2 when many create and delete operations are issued from more than one node in the same directory at the same time. If this causes performance problems in your system, you should localize file creation and deletions by a node to directories specific to that node as much as possible.
1.3. GFS2 support considerations
To be eligible for support from Red Hat for a cluster running a GFS2 file system, you must take into account the support policies for GFS2 file systems.
For full information about Red Hat’s support policies, requirements, and limitations for RHEL High Availability clusters, see Support Policies for RHEL High Availability Clusters.
1.3.1. Maximum file system and cluster size
The following table summarizes the current maximum file system size and number of nodes that GFS2 supports.
Parameter | Maximum |
---|---|
Number of nodes | 16 (x86, Power8 on PowerVM) 4 (s390x under z/VM) |
File system size | 100TB on all supported architectures |
GFS2 is based on a 64-bit architecture, which can theoretically accommodate an 8 EB file system. If your system requires larger GFS2 file systems than are currently supported, contact your Red Hat service representative.
When determining the size of your file system, you should consider your recovery needs. Running the fsck.gfs2
command on a very large file system can take a long time and consume a large amount of memory. Additionally, in the event of a disk or disk subsystem failure, recovery time is limited by the speed of your backup media. For information about the amount of memory the fsck.gfs2
command requires, see Determining required memory for running fsck.gfs2.
1.3.2. Minimum cluster size
Although a GFS2 file system can be implemented in a standalone system or as part of a cluster configuration, Red Hat does not support the use of GFS2 as a single-node file system, with the following exceptions:
- Red Hat supports single-node GFS2 file systems for mounting snapshots of cluster file systems as might be needed, for example, for backup purposes.
A single-node cluster mounting GFS2 file systems (which uses DLM) is supported for the purposes of a secondary-site Disaster Recovery (DR) node. This exception is for DR purposes only and not for transferring the main cluster workload to the secondary site.
For example, copying off the data from the filesystem mounted on the secondary site while the primary site is offline is supported. However, migrating a workload from the primary site directly to a single-node cluster secondary site is unsupported. If the full work load needs to be migrated to the single-node secondary site then the secondary site must be the same size as the primary site.
Red Hat recommends that when you mount a GFS2 file system in a single-node cluster you specify the
errors=panic
mount option so that the single-node cluster will panic when a GFS2 withdraw occurs since the single-node cluster will not be able to fence itself when encountering file system errors.
Red Hat supports a number of high-performance single-node file systems that are optimized for single node and thus have generally lower overhead than a cluster file system. Red Hat recommends using these file systems in preference to GFS2 in cases where only a single node needs to mount the file system. For information about the file systems that Red Hat Enterprise Linux 9 supports, see Managing file systems.
1.4. GFS2 formatting considerations
To format your GFS2 file system to optimize performance, you should take these recommendations into account.
Make sure that your deployment of the Red Hat High Availability Add-On meets your needs and can be supported. Consult with an authorized Red Hat representative to verify your configuration prior to deployment.
File System Size: Smaller Is Better
GFS2 is based on a 64-bit architecture, which can theoretically accommodate an 8 EB file system. However, the current supported maximum size of a GFS2 file system for 64-bit hardware is 100TB.
Note that even though GFS2 large file systems are possible, that does not mean they are recommended. The rule of thumb with GFS2 is that smaller is better: it is better to have 10 1TB file systems than one 10TB file system.
There are several reasons why you should keep your GFS2 file systems small:
- Less time is required to back up each file system.
-
Less time is required if you need to check the file system with the
fsck.gfs2
command. -
Less memory is required if you need to check the file system with the
fsck.gfs2
command.
In addition, fewer resource groups to maintain mean better performance.
Of course, if you make your GFS2 file system too small, you might run out of space, and that has its own consequences. You should consider your own use cases before deciding on a size.
Block Size: Default (4K) Blocks Are Preferred
The mkfs.gfs2
command attempts to estimate an optimal block size based on device topology. In general, 4K blocks are the preferred block size because 4K is the default page size (memory) for Red Hat Enterprise Linux. Unlike some other file systems, GFS2 does most of its operations using 4K kernel buffers. If your block size is 4K, the kernel has to do less work to manipulate the buffers.
It is recommended that you use the default block size, which should yield the highest performance. You may need to use a different block size only if you require efficient storage of many very small files.
Journal Size: Default (128MB) Is Usually Optimal
When you run the mkfs.gfs2
command to create a GFS2 file system, you may specify the size of the journals. If you do not specify a size, it will default to 128MB, which should be optimal for most applications.
Some system administrators might think that 128MB is excessive and be tempted to reduce the size of the journal to the minimum of 8MB or a more conservative 32MB. While that might work, it can severely impact performance. Like many journaling file systems, every time GFS2 writes metadata, the metadata is committed to the journal before it is put into place. This ensures that if the system crashes or loses power, you will recover all of the metadata when the journal is automatically replayed at mount time. However, it does not take much file system activity to fill an 8MB journal, and when the journal is full, performance slows because GFS2 has to wait for writes to the storage.
It is generally recommended to use the default journal size of 128MB. If your file system is very small (for example, 5GB), having a 128MB journal might be impractical. If you have a larger file system and can afford the space, using 256MB journals might improve performance.
Size and Number of Resource Groups
When a GFS2 file system is created with the mkfs.gfs2
command, it divides the storage into uniform slices known as resource groups. It attempts to estimate an optimal resource group size (ranging from 32MB to 2GB). You can override the default with the -r
option of the mkfs.gfs2
command.
Your optimal resource group size depends on how you will use the file system. Consider how full it will be and whether or not it will be severely fragmented.
You should experiment with different resource group sizes to see which results in optimal performance. It is a best practice to experiment with a test cluster before deploying GFS2 into full production.
If your file system has too many resource groups, each of which is too small, block allocations can waste too much time searching tens of thousands of resource groups for a free block. The more full your file system, the more resource groups that will be searched, and every one of them requires a cluster-wide lock. This leads to slow performance.
If, however, your file system has too few resource groups, each of which is too big, block allocations might contend more often for the same resource group lock, which also impacts performance. For example, if you have a 10GB file system that is carved up into five resource groups of 2GB, the nodes in your cluster will fight over those five resource groups more often than if the same file system were carved into 320 resource groups of 32MB. The problem is exacerbated if your file system is nearly full because every block allocation might have to look through several resource groups before it finds one with a free block. GFS2 tries to mitigate this problem in two ways:
- First, when a resource group is completely full, it remembers that and tries to avoid checking it for future allocations until a block is freed from it. If you never delete files, contention will be less severe. However, if your application is constantly deleting blocks and allocating new blocks on a file system that is mostly full, contention will be very high and this will severely impact performance.
- Second, when new blocks are added to an existing file (for example, by appending) GFS2 will attempt to group the new blocks together in the same resource group as the file. This is done to increase performance: on a spinning disk, seek operations take less time when they are physically close together.
The worst case scenario is when there is a central directory in which all the nodes create files because all of the nodes will constantly fight to lock the same resource group.
1.5. Considerations for GFS2 in a cluster
When determining the number of nodes that your system will contain, note that there is a trade-off between high availability and performance. With a larger number of nodes, it becomes increasingly difficult to make workloads scale. For that reason, Red Hat does not support using GFS2 for cluster file system deployments greater than 16 nodes.
Deploying a cluster file system is not a "drop in" replacement for a single node deployment. Red Hat recommends that you allow a period of around 8-12 weeks of testing on new installations in order to test the system and ensure that it is working at the required performance level. During this period, any performance or functional issues can be worked out and any queries should be directed to the Red Hat support team.
Red Hat recommends that customers considering deploying clusters have their configurations reviewed by Red Hat support before deployment to avoid any possible support issues later on.
1.6. Hardware considerations
Take the following hardware considerations into account when deploying a GFS2 file system.
Use higher quality storage options
GFS2 can operate on cheaper shared storage options, such as iSCSI or Fibre Channel over Ethernet (FCoE), but you will get better performance if you buy higher quality storage with larger caching capacity. Red Hat performs most quality, sanity, and performance tests on SAN storage with Fibre Channel interconnect. As a general rule, it is always better to deploy something that has been tested first.
Test network equipment before deploying
Higher quality, faster network equipment makes cluster communications and GFS2 run faster with better reliability. However, you do not have to purchase the most expensive hardware. Some of the most expensive network switches have problems passing multicast packets, which are used for passing
fcntl
locks (flocks), whereas cheaper commodity network switches are sometimes faster and more reliable. Red Hat recommends trying equipment before deploying it into full production.
Chapter 2. Recommendations for GFS2 usage
When deploying a GFS2 file system, there are a variety of general recommendations you should take into account.
2.1. Configuring atime
updates
Each file inode and directory inode has three time stamps associated with it:
-
ctime
— The last time the inode status was changed -
mtime
— The last time the file (or directory) data was modified -
atime
— The last time the file (or directory) data was accessed
If atime
updates are enabled as they are by default on GFS2 and other Linux file systems, then every time a file is read its inode needs to be updated.
Because few applications use the information provided by atime
, those updates can require a significant amount of unnecessary write traffic and file locking traffic. That traffic can degrade performance; therefore, it may be preferable to turn off or reduce the frequency of atime
updates.
The following methods of reducing the effects of atime
updating are available:
-
Mount with
relatime
(relative atime), which updates theatime
if the previousatime
update is older than themtime
orctime
update. This is the default mount option for GFS2 file systems. -
Mount with
noatime
ornodiratime
. Mounting withnoatime
disablesatime
updates for both files and directories on that file system, while mounting withnodiratime
disablesatime
updates only for directories on that file system, It is generally recommended that you mount GFS2 file systems with thenoatime
ornodiratime
mount option whenever possible, with the preference fornoatime
where the application allows for this. For more information about the effect of these arguments on GFS2 file system performance, see GFS2 Node Locking.
Use the following command to mount a GFS2 file system with the noatime
Linux mount option.
mount BlockDevice MountPoint -o noatime
BlockDevice
- Specifies the block device where the GFS2 file system resides.
MountPoint
- Specifies the directory where the GFS2 file system should be mounted.
In this example, the GFS2 file system resides on /dev/vg01/lvol0
and is mounted on directory /mygfs2
with atime
updates turned off.
# mount /dev/vg01/lvol0 /mygfs2 -o noatime
2.2. VFS tuning options: research and experiment
Like all Linux file systems, GFS2 sits on top of a layer called the virtual file system (VFS). The VFS provides good defaults for the cache settings for most workloads and should not need changing in most cases. If, however, you have a workload that is not running efficiently (for example, cache is too large or too small) then you may be able to improve the performance by using the sysctl
(8) command to adjust the values of the sysctl
files in the /proc/sys/vm
directory. Documentation for these files can be found in the kernel source tree Documentation/sysctl/vm.txt
.
For example, the values for dirty_background_ratio
and vfs_cache_pressure
may be adjusted depending on your situation. To fetch the current values, use the following commands:
# sysctl -n vm.dirty_background_ratio # sysctl -n vm.vfs_cache_pressure
The following commands adjust the values:
# sysctl -w vm.dirty_background_ratio=20 # sysctl -w vm.vfs_cache_pressure=500
You can permanently change the values of these parameters by editing the /etc/sysctl.conf
file.
To find the optimal values for your use cases, research the various VFS options and experiment on a test cluster before deploying into full production.
2.3. SELinux on GFS2
Use of Security Enhanced Linux (SELinux) with GFS2 incurs a small performance penalty. To avoid this overhead, you may choose not to use SELinux with GFS2 even on a system with SELinux in enforcing mode. When mounting a GFS2 file system, you can ensure that SELinux will not attempt to read the seclabel
element on each file system object by using one of the context
options as described on the mount
(8) man page; SELinux will assume that all content in the file system is labeled with the seclabel
element provided in the context
mount options. This will also speed up processing as it avoids another disk read of the extended attribute block that could contain seclabel
elements.
For example, on a system with SELinux in enforcing mode, you can use the following mount
command to mount the GFS2 file system if the file system is going to contain Apache content. This label will apply to the entire file system; it remains in memory and is not written to disk.
# mount -t gfs2 -o context=system_u:object_r:httpd_sys_content_t:s0 /dev/mapper/xyz/mnt/gfs2
If you are not sure whether the file system will contain Apache content, you can use the labels public_content_rw_t
or public_content_t
, or you could define a new label altogether and define a policy around it.
Note that in a Pacemaker cluster you should always use Pacemaker to manage a GFS2 file system. You can specify the mount options when you create a GFS2 file system resource.
2.4. Setting up NFS over GFS2
Due to the added complexity of the GFS2 locking subsystem and its clustered nature, setting up NFS over GFS2 requires taking many precautions.
If the GFS2 file system is NFS exported, then you must mount the file system with the localflocks
option. Because utilizing the localflocks
option prevents you from safely accessing the GFS2 filesystem from multiple locations, and it is not viable to export GFS2 from multiple nodes simultaneously, it is a support requirement that the GFS2 file system be mounted on only one node at a time when using this configuration. The intended effect of this is to force POSIX locks from each server to be local: non-clustered, independent of each other. This is because a number of problems exist if GFS2 attempts to implement POSIX locks from NFS across the nodes of a cluster. For applications running on NFS clients, localized POSIX locks means that two clients can hold the same lock concurrently if the two clients are mounting from different servers, which could cause data corruption. If all clients mount NFS from one server, then the problem of separate servers granting the same locks independently goes away. If you are not sure whether to mount your file system with the localflocks
option, you should not use the option. Contact Red Hat support immediately to discuss the appropriate configuration to avoid data loss. Exporting GFS2 via NFS, while technically supported in some circumstances, is not recommended.
For all other (non-NFS) GFS2 applications, do not mount your file system using localflocks
, so that GFS2 will manage the POSIX locks and flocks between all the nodes in the cluster (on a cluster-wide basis). If you specify localflocks
and do not use NFS, the other nodes in the cluster will not have knowledge of each other’s POSIX locks and flocks, thus making them unsafe in a clustered environment
In addition to the locking considerations, you should take the following into account when configuring an NFS service over a GFS2 file system.
Red Hat supports only Red Hat High Availability Add-On configurations using NFSv3 with locking in an active/passive configuration with the following characteristics. This configuration provides High Availability (HA) for the file system and reduces system downtime since a failed node does not result in the requirement to execute the
fsck
command when failing the NFS server from one node to another.- The back-end file system is a GFS2 file system running on a 2 to 16 node cluster.
- An NFSv3 server is defined as a service exporting the entire GFS2 file system from a single cluster node at a time.
- The NFS server can fail over from one cluster node to another (active/passive configuration).
- No access to the GFS2 file system is allowed except through the NFS server. This includes both local GFS2 file system access as well as access through Samba or Clustered Samba. Accessing the file system locally via the cluster node from which it is mounted may result in data corruption.
- There is no NFS quota support on the system.
-
The
fsid=
NFS option is mandatory for NFS exports of GFS2. - If problems arise with your cluster (for example, the cluster becomes inquorate and fencing is not successful), the clustered logical volumes and the GFS2 file system will be frozen and no access is possible until the cluster is quorate. You should consider this possibility when determining whether a simple failover solution such as the one defined in this procedure is the most appropriate for your system.
2.5. Samba (SMB or Windows) file serving over GFS2
You can use Samba (SMB or Windows) file serving from a GFS2 file system with CTDB, which allows active/active configurations.
Simultaneous access to the data in the Samba share from outside of Samba is not supported. There is currently no support for GFS2 cluster leases, which slows Samba file serving. For further information about support policies for Samba, see Support Policies for RHEL Resilient Storage - ctdb General Policies and Support Policies for RHEL Resilient Storage - Exporting gfs2 contents via other protocols.
2.6. Configuring virtual machines for GFS2
When using a GFS2 file system with a virtual machine, it is important that your VM storage settings on each node be configured properly in order to force the cache off. For example, including these settings for cache
and io
in the libvirt
domain should allow GFS2 to behave as expected.
<driver name='qemu' type='raw' cache='none' io='native'/>
Alternately, you can configure the shareable
attribute within the device element. This indicates that the device is expected to be shared between domains (as long as hypervisor and OS support this). If shareable
is used, cache='no'
should be used for that device.
2.7. Block allocation
Even though applications that only write data typically do not care how or where a block is allocated, some knowledge of how block allocation works can help you optimize performance.
2.7.1. Leave free space in the file system
When a GFS2 file system is nearly full, the block allocator starts to have a difficult time finding space for new blocks to be allocated. As a result, blocks given out by the allocator tend to be squeezed into the end of a resource group or in tiny slices where file fragmentation is much more likely. This file fragmentation can cause performance problems. In addition, when a GFS2 file system is nearly full, the GFS2 block allocator spends more time searching through multiple resource groups, and that adds lock contention that would not necessarily be there on a file system that has ample free space. This also can cause performance problems.
For these reasons, it is recommended that you not run a file system that is more than 85 percent full, although this figure may vary depending on workload.
2.7.2. Have each node allocate its own files, if possible
When developing applications for use with GFS2 file systems, it is recommended that you have each node allocate it own files, if possible. Due to the way the distributed lock manager (DLM) works, there will be more lock contention if all files are allocated by one node and other nodes need to add blocks to those files.
The term "lock master" has been used historically to denote a node which is currently the coordinator of lock requests, which originate locally or from a remote node in the cluster. This term for the lock request coordinator is slightly misleading because it is really a resource (in DLM terminology) in relation to which lock requests are either queued, granted or declined. In the sense in which the term is used in the DLM, it should be taken to refer to "first among equals", since the DLM is a peer-to-peer system.
In the Linux kernel DLM implementation, the node on which the lock is first used becomes the coordinator of lock requests, and after that point it does not change. This is an implementation detail of the Linux kernel DLM and not a property of DLMs in general. It is possible that a future update may allow the coordination of lock requests for a particular lock to move between nodes.
The location where lock requests are coordinated is transparent to the initiator of the lock request, except by the effect on the latency of the request. One consequence of the current implementation is that if there is an imbalance of the initial workload (for example, one node scans through the whole filesystem before others perform any I/O commands) this can result in higher lock latencies for other nodes in the cluster compared with the node that performed the initial scan of the filesystem.
As in many file systems, the GFS2 allocator tries to keep blocks in the same file close to one another to reduce the movement of disk heads and boost performance. A node that allocates blocks to a file will likely need to use and lock the same resource groups for the new blocks (unless all the blocks in that resource group are in use). The file system will run faster if the lock request coordinator for the resource group containing the file allocates its data blocks (it is faster to have the node that first opened the file do all the writing of new blocks).
2.7.3. Preallocate, if possible
If files are preallocated, block allocations can be avoided altogether and the file system can run more efficiently. GFS2 includes the fallocate(1)
system call, which you can use to preallocate blocks of data.
Chapter 3. Administering GFS2 file systems
There are a variety of commands and options that you use to create, mount, grow, and manage GFS2 file systems.
3.1. GFS2 file system creation
You create a GFS2 file system with the mkfs.gfs2
command. A file system is created on an activated LVM volume.
3.1.1. The GFS2 mkfs command
The following information is required to run the mkfs.gfs2
command to create a clustered GFS2 file system:
-
Lock protocol/module name, which is
lock_dlm
for a cluster - Cluster name
- Number of journals (one journal required for each node that may be mounting the file system)
Once you have created a GFS2 file system with the mkfs.gfs2
command, you cannot decrease the size of the file system. You can, however, increase the size of an existing file system with the gfs2_grow
command.
The format for creating a clustered GFS2 file system is as follows. Note that Red Hat does not support the use of GFS2 as a single-node file system.
mkfs.gfs2 -p lock_dlm -t ClusterName:FSName -j NumberJournals BlockDevice
If you prefer, you can create a GFS2 file system by using the mkfs
command with the -t
parameter specifying a file system of type gfs2
, followed by the GFS2 file system options.
mkfs -t gfs2 -p lock_dlm -t ClusterName:FSName -j NumberJournals BlockDevice
Improperly specifying the ClusterName:FSName parameter may cause file system or lock space corruption.
ClusterName
- The name of the cluster for which the GFS2 file system is being created.
FSName
-
The file system name, which can be 1 to 16 characters long. The name must be unique for all
lock_dlm
file systems over the cluster. NumberJournals
-
Specifies the number of journals to be created by the
mkfs.gfs2
command. One journal is required for each node that mounts the file system. For GFS2 file systems, more journals can be added later without growing the file system. BlockDevice
- Specifies a logical or other block device
The following table describes the mkfs.gfs2
command options (flags and parameters).
Flag | Parameter | Description |
---|---|---|
|
|
Sets the initial size of each journal’s quota change file to |
| Enables debugging output. | |
| Help. Displays available options. | |
|
| Specifies the size of the journal in megabytes. Default journal size is 128 megabytes. The minimum size is 8 megabytes. Larger journals improve performance, although they use more memory than smaller journals. |
|
|
Specifies the number of journals to be created by the |
|
Prevents the | |
|
| * Specifies the name of the locking protocol to use. Recognized locking protocols include:
*
* |
| Quiet. Do not display anything. | |
|
|
Specifies the size of the resource groups in megabytes. The minimum resource group size is 32 megabytes. The maximum resource group size is 2048 megabytes. A large resource group size may increase performance on very large file systems. If this is not specified, |
|
|
* A unique identifier that specifies the lock table field when you use the
* This parameter has two parts separated by a colon (no spaces) as follows:
*
* |
| Displays command version information. |
3.1.2. Creating a GFS2 file system
The following example creates two GFS2 file systems. For both of these file systems, lock_dlm` is the locking protocol that the file system uses, since this is a clustered file system. Both file systems can be used in the cluster named alpha
.
For the first file system, file system name is mydata1
. it contains eight journals and is created on /dev/vg01/lvol0
. For the second file system, the file system name is mydata2
. It contains eight journals and is created on /dev/vg01/lvol1
.
# mkfs.gfs2 -p lock_dlm -t alpha:mydata1 -j 8 /dev/vg01/lvol0 # mkfs.gfs2 -p lock_dlm -t alpha:mydata2 -j 8 /dev/vg01/lvol1
3.2. Mounting a GFS2 file system
Before you can mount a GFS2 file system, the file system must exist, the volume where the file system exists must be activated, and the supporting clustering and locking systems must be started. After those requirements have been met, you can mount the GFS2 file system as you would any Linux file system.
You should always use Pacemaker to manage the GFS2 file system in a production environment rather than manually mounting the file system with a mount
command, as this may cause issues at system shutdown.
To manipulate file ACLs, you must mount the file system with the -o acl
mount option. If a file system is mounted without the -o acl
mount option, users are allowed to view ACLs (with getfacl
), but are not allowed to set them (with setfacl
).
3.2.1. Mounting a GFS2 file system with no options specified
In this example, the GFS2 file system on /dev/vg01/lvol0
is mounted on the /mygfs2
directory.
# mount /dev/vg01/lvol0 /mygfs2
3.2.2. Mounting a GFS2 file system that specifies mount options
The following is the format for the command to mount a GFS2 file system that specifies mount options.
mount BlockDevice MountPoint -o option
BlockDevice
- Specifies the block device where the GFS2 file system resides.
MountPoint
- Specifies the directory where the GFS2 file system should be mounted.
The -o option
argument consists of GFS2-specific options or acceptable standard Linux mount -o
options, or a combination of both. Multiple option
parameters are separated by a comma and no spaces.
The mount
command is a Linux system command. In addition to using these GFS2-specific options, you can use other, standard, mount
command options (for example, -r
). For information about other Linux mount
command options, see the Linux mount
man page.
The following table describes the available GFS2-specific -o option
values that can be passed to GFS2 at mount time.
This table includes descriptions of options that are used with local file systems only. Note, however, that Red Hat does not support the use of GFS2 as a single-node file system. Red Hat will continue to support single-node GFS2 file systems for mounting snapshots of cluster file systems (for example, for backup purposes).
Option | Description |
---|---|
|
Allows manipulating file ACLs. If a file system is mounted without the |
|
When |
|
Forces GFS2 to treat the file system as a multi-host file system. By default, using |
|
Tells GFS2 to let the VFS (virtual file system) layer do all flock and fcntl. The |
|
Allows the user to specify which locking protocol to use with the file system. If |
| Allows the user to specify which locking table to use with the file system. |
|
Turns quotas on or off for a file system. Setting the quotas to be in the |
|
When |
| Causes GFS2 to generate "discard" I/O requests for blocks that have been freed. These can be used by suitable hardware to implement thin provisioning and similar schemes. |
|
Causes GFS2 to send I/O barriers when flushing the journal. The default value is |
| Sets the number of seconds for which a change in the quota information may sit on one node before being written to the quota file. This is the preferred way to set this parameter. The value is an integer number of seconds greater than zero. The default is 60 seconds. Shorter settings result in faster updates of the lazy quota information and less likelihood of someone exceeding their quota. Longer settings make file system operations involving quotas faster and more efficient. |
|
Setting |
|
Provides a bound on the maximum percentage change in the |
3.2.3. Unmounting a GFS2 file system
GFS2 file systems that have been mounted manually rather than automatically through Pacemaker will not be known to the system when file systems are unmounted at system shutdown. As a result, the GFS2 resource agent will not unmount the GFS2 file system. After the GFS2 resource agent is shut down, the standard shutdown process kills off all remaining user processes, including the cluster infrastructure, and tries to unmount the file system. This unmount will fail without the cluster infrastructure and the system will hang.
To prevent the system from hanging when the GFS2 file systems are unmounted, you should do one of the following:
- Always use Pacemaker to manage the GFS2 file system.
-
If a GFS2 file system has been mounted manually with the
mount
command, be sure to unmount the file system manually with theumount
command before rebooting or shutting down the system.
If your file system hangs while it is being unmounted during system shutdown under these circumstances, perform a hardware reboot. It is unlikely that any data will be lost since the file system is synced earlier in the shutdown process.
The GFS2 file system can be unmounted the same way as any Linux file system, by using the umount
command.
The umount
command is a Linux system command. Information about this command can be found in the Linux umount
command man pages.
Usage
umount MountPoint
MountPoint
- Specifies the directory where the GFS2 file system is currently mounted.
3.3. Backing up a GFS2 file system
It is important to make regular backups of your GFS2 file system in case of emergency, regardless of the size of your file system. Many system administrators feel safe because they are protected by RAID, multipath, mirroring, snapshots, and other forms of redundancy, but there is no such thing as safe enough.
It can be a problem to create a backup since the process of backing up a node or set of nodes usually involves reading the entire file system in sequence. If this is done from a single node, that node will retain all the information in cache until other nodes in the cluster start requesting locks. Running this type of backup program while the cluster is in operation will negatively impact performance.
Dropping the caches once the backup is complete reduces the time required by other nodes to regain ownership of their cluster locks and caches. This is still not ideal, however, because the other nodes will have stopped caching the data that they were caching before the backup process began. You can drop caches using the following command after the backup is complete:
echo -n 3 > /proc/sys/vm/drop_caches
It is faster if each node in the cluster backs up its own files so that the task is split between the nodes. You might be able to accomplish this with a script that uses the rsync
command on node-specific directories.
Red Hat recommends making a GFS2 backup by creating a hardware snapshot on the SAN, presenting the snapshot to another system, and backing it up there. The backup system should mount the snapshot with -o lockproto=lock_nolock
since it will not be in a cluster. Note, however, that Red Hat does not support the use of GFS2 as a single-node file system in a production environment. This option should be used only for the purposes of backup or for a secondary-site Disaster Recovery node, as described in Minimum cluster size. When using this option, you must ensure that the GFS2 file system is being used by only one system at a time.
3.4. Suspending activity on a GFS2 file system
You can suspend write activity to a file system by using the dmsetup suspend
command. Suspending write activity allows hardware-based device snapshots to be used to capture the file system in a consistent state. The dmsetup resume
command ends the suspension.
The format for the command to suspend activity on a GFS2 file system is as follows.
dmsetup suspend MountPoint
This example suspends writes to file system /mygfs2
.
# dmsetup suspend /mygfs2
The format for the command to end suspension of activity on a GFS2 file system is as follows.
dmsetup resume MountPoint
This example ends suspension of writes to file system /mygfs2
.
# dmsetup resume /mygfs2
3.5. Growing a GFS2 file system
The gfs2_grow
command is used to expand a GFS2 file system after the device where the file system resides has been expanded. Running the gfs2_grow
command on an existing GFS2 file system fills all spare space between the current end of the file system and the end of the device with a newly initialized GFS2 file system extension. All nodes in the cluster can then use the extra storage space that has been added.
You cannot decrease the size of a GFS2 file system.
The gfs2_grow
command must be run on a mounted file system. The following procedure increases the size of the GFS2 file system in a cluster that is mounted on the logical volume shared_vg/shared_lv1
with a mount point of /mnt/gfs2
.
Procedure
- Perform a backup of the data on the file system.
If you do not know the logical volume that is used by the file system to be expanded, you can determine this by running the
df mountpoint
command. This will display the device name in the following format:/dev/mapper/vg-lv
For example, the device name
/dev/mapper/shared_vg-shared_lv1
indicates that the logical volume isshared_vg/shared_lv1
.On one node of the cluster, expand the underlying cluster volume with the
lvextend
command.# lvextend -L+1G shared_vg/shared_lv1 Size of logical volume shared_vg/shared_lv1 changed from 5.00 GiB (1280 extents) to 6.00 GiB (1536 extents). WARNING: extending LV with a shared lock, other hosts may require LV refresh. Logical volume shared_vg/shared_lv1 successfully resized.
One one node of the cluster, increase the size of the GFS2 file system. Do not extend the file system if the logical volume was not refreshed on all of the nodes, otherwise the file system data may become unavailable throughout the cluster.
# gfs2_grow /mnt/gfs2 FS: Mount point: /mnt/gfs2 FS: Device: /dev/mapper/shared_vg-shared_lv1 FS: Size: 1310719 (0x13ffff) DEV: Length: 1572864 (0x180000) The file system will grow by 1024MB. gfs2_grow complete.
Run the
df
command on all nodes to check that the new space is now available in the file system. Note that it may take up to 30 seconds for thedf
command on all nodes to show the same file system size# df -h /mnt/gfs2] Filesystem Size Used Avail Use% Mounted on /dev/mapper/shared_vg-shared_lv1 6.0G 4.5G 1.6G 75% /mnt/gfs2
3.6. Adding journals to a GFS2 file system
GFS2 requires one journal for each node in a cluster that needs to mount the file system. If you add additional nodes to the cluster, you can add journals to a GFS2 file system with the gfs2_jadd
command. You can add journals to a GFS2 file system dynamically at any point without expanding the underlying logical volume. The gfs2_jadd
command must be run on a mounted file system, but it needs to be run on only one node in the cluster. All the other nodes sense that the expansion has occurred.
If a GFS2 file system is full, the gfs2_jadd
command will fail, even if the logical volume containing the file system has been extended and is larger than the file system. This is because in a GFS2 file system, journals are plain files rather than embedded metadata, so simply extending the underlying logical volume will not provide space for the journals.
Before adding journals to a GFS2 file system, you can find out how many journals the GFS2 file system currently contains with the gfs2_edit -p jindex
command, as in the following example:
# gfs2_edit -p jindex /dev/sasdrives/scratch|grep journal
3/3 [fc7745eb] 4/25 (0x4/0x19): File journal0
4/4 [8b70757d] 5/32859 (0x5/0x805b): File journal1
5/5 [127924c7] 6/65701 (0x6/0x100a5): File journal2
The format for the basic command to add journals to a GFS2 file system is as follows.
gfs2_jadd -j Number MountPoint
Number
- Specifies the number of new journals to be added.
MountPoint
- Specifies the directory where the GFS2 file system is mounted.
In this example, one journal is added to the file system on the /mygfs2
directory.
# gfs2_jadd -j 1 /mygfs2
Chapter 4. GFS2 quota management
File system quotas are used to limit the amount of file system space a user or group can use. A user or group does not have a quota limit until one is set. When a GFS2 file system is mounted with the quota=on
or quota=account
option, GFS2 keeps track of the space used by each user and group even when there are no limits in place. GFS2 updates quota information in a transactional way so system crashes do not require quota usages to be reconstructed. To prevent a performance slowdown, a GFS2 node synchronizes updates to the quota file only periodically. The fuzzy quota accounting can allow users or groups to slightly exceed the set limit. To minimize this, GFS2 dynamically reduces the synchronization period as a hard quota limit is approached.
GFS2 supports the standard Linux quota facilities. In order to use this you will need to install the quota RPM. This is the preferred way to administer quotas on GFS2 and should be used for all new deployments of GFS2 using quotas.
For more information about disk quotas, see the man
pages of the following commands:
-
quotacheck
-
edquota
-
repquota
-
quota
4.1. Configuring GFS2 disk quotas
To implement disk quotas for GFS2 file systems, there are three steps to perform.
The steps to perform to implement disk quotas are as follows:
- Set up quotas in enforcement or accounting mode.
- Initialize the quota database file with current block usage information.
- Assign quota policies. (In accounting mode, these policies are not enforced.)
Each of these steps is discussed in detail in the following sections.
4.1.1. Setting up quotas in enforcement or accounting mode
In GFS2 file systems, quotas are disabled by default. To enable quotas for a file system, mount the file system with the quota=on
option specified.
To mount a file system with quotas enabled, specify quota=on
for the options
argument when creating the GFS2 file system resource in a cluster. For example, the following command specifies that the GFS2 Filesystem
resource being created will be mounted with quotas enabled.
# pcs resource create gfs2mount Filesystem options="quota=on" device=BLOCKDEVICE directory=MOUNTPOINT fstype=gfs2 clone
It is possible to keep track of disk usage and maintain quota accounting for every user and group without enforcing the limit and warn values. To do this, mount the file system with the quota=account
option specified.
To mount a file system with quotas disabled, specify quota=off
for the options
argument when creating the GFS2 file system resource in a cluster.
4.1.2. Creating the quota database files
After each quota-enabled file system is mounted, the system is capable of working with disk quotas. However, the file system itself is not yet ready to support quotas. The next step is to run the quotacheck
command.
The quotacheck
command examines quota-enabled file systems and builds a table of the current disk usage per file system. The table is then used to update the operating system’s copy of disk usage. In addition, the file system’s disk quota files are updated.
To create the quota files on the file system, use the -u
and the -g
options of the quotacheck
command; both of these options must be specified for user and group quotas to be initialized. For example, if quotas are enabled for the /home
file system, create the files in the /home
directory:
# quotacheck -ug /home
4.1.3. Assigning quotas per user
The last step is assigning the disk quotas with the edquota
command. Note that if you have mounted your file system in accounting mode (with the quota=account
option specified), the quotas are not enforced.
To configure the quota for a user, as root in a shell prompt, execute the command:
# edquota username
Perform this step for each user who needs a quota. For example, if a quota is enabled for the /home
partition (/dev/VolGroup00/LogVol02
in the example below) and the command edquota testuser
is executed, the following is shown in the editor configured as the default for the system:
Disk quotas for user testuser (uid 501): Filesystem blocks soft hard inodes soft hard /dev/VolGroup00/LogVol02 440436 0 0
The text editor defined by the EDITOR
environment variable is used by edquota
. To change the editor, set the EDITOR
environment variable in your ~/.bash_profile
file to the full path of the editor of your choice.
The first column is the name of the file system that has a quota enabled for it. The second column shows how many blocks the user is currently using. The next two columns are used to set soft and hard block limits for the user on the file system.
The soft block limit defines the maximum amount of disk space that can be used.
The hard block limit is the absolute maximum amount of disk space that a user or group can use. Once this limit is reached, no further disk space can be used.
The GFS2 file system does not maintain quotas for inodes, so these columns do not apply to GFS2 file systems and will be blank.
If any of the values are set to 0, that limit is not set. In the text editor, change the limits. For example:
Disk quotas for user testuser (uid 501): Filesystem blocks soft hard inodes soft hard /dev/VolGroup00/LogVol02 440436 500000 550000
To verify that the quota for the user has been set, use the following command:
# quota testuser
You can also set quotas from the command line with the setquota
command. For information about the setquota
command, see the setquota
(8) man page.
4.1.4. Assigning quotas per group
Quotas can also be assigned on a per-group basis. Note that if you have mounted your file system in accounting mode (with the account=on
option specified), the quotas are not enforced.
To set a group quota for the devel
group (the group must exist prior to setting the group quota), use the following command:
# edquota -g devel
This command displays the existing quota for the group in the text editor:
Disk quotas for group devel (gid 505): Filesystem blocks soft hard inodes soft hard /dev/VolGroup00/LogVol02 440400 0 0
The GFS2 file system does not maintain quotas for inodes, so these columns do not apply to GFS2 file systems and will be blank. Modify the limits, then save the file.
To verify that the group quota has been set, use the following command:
$ quota -g devel
4.2. Managing GFS2 disk Quotas
If quotas are implemented, they need some maintenance, mostly in the form of watching to see if the quotas are exceeded and making sure the quotas are accurate.
If users repeatedly exceed their quotas or consistently reach their soft limits, a system administrator has a few choices to make depending on what type of users they are and how much disk space impacts their work. The administrator can either help the user determine how to use less disk space or increase the user’s disk quota.
You can create a disk usage report by running the repquota
utility. For example, the command repquota /home
produces this output:
*** Report for user quotas on device /dev/mapper/VolGroup00-LogVol02 Block grace time: 7days; Inode grace time: 7days Block limits File limits User used soft hard grace used soft hard grace ---------------------------------------------------------------------- root -- 36 0 0 4 0 0 kristin -- 540 0 0 125 0 0 testuser -- 440400 500000 550000 37418 0 0
To view the disk usage report for all (option -a
) quota-enabled file systems, use the command:
# repquota -a
The --
displayed after each user is a quick way to determine whether the block limits have been exceeded. If the block soft limit is exceeded, a +
appears in place of the first -
in the output. The second -
indicates the inode limit, but GFS2 file systems do not support inode limits so that character will remain as -
. GFS2 file systems do not support a grace period, so the grace
column will remain blank.
Note that the repquota
command is not supported over NFS, irrespective of the underlying file system.
4.3. Keeping GFS2 disk quotas accurate with the quotacheck command
If you enable quotas on your file system after a period of time when you have been running with quotas disabled, you should run the quotacheck
command to create, check, and repair quota files. Additionally, you may want to run the quotacheck
command if you think your quota files may not be accurate, as may occur when a file system is not unmounted cleanly after a system crash.
For more information about the quotacheck
command, see the quotacheck(8)
man page.
Run quotacheck
when the file system is relatively idle on all nodes because disk activity may affect the computed quota values.
4.4. Synchronizing quotas with the quotasync Command
GFS2 stores all quota information in its own internal file on disk. A GFS2 node does not update this quota file for every file system write; rather, by default it updates the quota file once every 60 seconds. This is necessary to avoid contention among nodes writing to the quota file, which would cause a slowdown in performance.
As a user or group approaches their quota limit, GFS2 dynamically reduces the time between its quota-file updates to prevent the limit from being exceeded. The normal time period between quota synchronizations is a tunable parameter, quota_quantum
. You can change this from its default value of 60 seconds using the quota_quantum=
mount option, as described in the "GFS2-Specific Mount Options" table in Mounting a GFS2 file system that specifies mount options.
The quota_quantum
parameter must be set on each node and each time the file system is mounted. Changes to the quota_quantum
parameter are not persistent across unmounts. You can update the quota_quantum
value with the mount -o remount
.
You can use the quotasync
command to synchronize the quota information from a node to the on-disk quota file between the automatic updates performed by GFS2. Usage Synchronizing Quota Information
quotasync [-ug] -a|mountpoint...
u
- Sync the user quota files.
g
- Sync the group quota files
a
- Sync all file systems that are currently quota-enabled and support sync. When -a is absent, a file system mountpoint should be specified.
mountpoint
- Specifies the GFS2 file system to which the actions apply.
You can tune the time between synchronizations by specifying a quota-quantum
mount option.
# mount -o quota_quantum=secs,remount BlockDevice MountPoint
MountPoint
- Specifies the GFS2 file system to which the actions apply.
secs
- Specifies the new time period between regular quota-file synchronizations by GFS2. Smaller values may increase contention and slow down performance.
The following example synchronizes all the cached dirty quotas from the node it is run on to the on-disk quota file for the file system /mnt/mygfs2
.
# quotasync -ug /mnt/mygfs2
This following example changes the default time period between regular quota-file updates to one hour (3600 seconds) for file system /mnt/mygfs2
when remounting that file system on logical volume /dev/volgroup/logical_volume
.
# mount -o quota_quantum=3600,remount /dev/volgroup/logical_volume /mnt/mygfs2
Chapter 5. GFS2 file system repair
When nodes fail with the file system mounted, file system journaling allows fast recovery. However, if a storage device loses power or is physically disconnected, file system corruption may occur. (Journaling cannot be used to recover from storage subsystem failures.) When that type of corruption occurs, you can recover the GFS2 file system by using the fsck.gfs2
command.
The fsck.gfs2
command must be run only on a file system that is unmounted from all nodes. When the file system is being managed as a Pacemaker cluster resource, you can disable the file system resource, which unmounts the file system. After running the fsck.gfs2
command, you enable the file system resource again. The timeout value specified with the --wait
option of the pcs resource disable
indicates a value in seconds.
pcs resource disable --wait=timeoutvalue resource_id [fsck.gfs2] pcs resource enable resource_id
Note that even if a file system is part of a resource group, as in an encrypted file system deployment, you need to disable only the file system resource in order to run the fsck command on the file system. You must not disable the entire resource group.
To ensure that fsck.gfs2
command does not run on a GFS2 file system at boot time, you can set the run_fsck
parameter of the options
argument when creating the GFS2 file system resource in a cluster. Specifying "run_fsck=no"
will indicate that you should not run the fsck
command.
5.1. Determining required memory for running fsck.gfs2
Running the fsck.gfs2
command may require system memory above and beyond the memory used for the operating system and kernel. Larger file systems in particular may require additional memory to run this command.
The following table shows approximate values of memory that may be required to run fsck.gfs2
file systems on GFS2 file systems that are 1TB, 10TB, and 100TB in size with a block size of 4K.
GFS2 file system size | Approximate memory required to run fsck.gfs2 |
---|---|
1 TB | 0.16 GB |
10 TB | 1.6 GB |
100 TB | 16 GB |
Note that a smaller block size for the file system would require a larger amount of memory. For example, GFS2 file systems with a block size of 1K would require four times the amount of memory indicated in this table.
5.2. Repairing a gfs2 filesystem
The format of the fsck.gfs2
command to repair a GFS2 filesystem is as follows:
fsck.gfs2 -y BlockDevice
-y
-
The
-y
flag causes all questions to be answered withyes
. With the-y
flag specified, thefsck.gfs2
command does not prompt you for an answer before making changes. BlockDevice
- Specifies the block device where the GFS2 file system resides.
In this example, the GFS2 file system residing on block device /dev/testvg/testlv
is repaired. All queries to repair are automatically answered with yes
.
# fsck.gfs2 -y /dev/testvg/testlv
Initializing fsck
Validating Resource Group index.
Level 1 RG check.
(level 1 passed)
Clearing journals (this may take a while)...
Journals cleared.
Starting pass1
Pass1 complete
Starting pass1b
Pass1b complete
Starting pass1c
Pass1c complete
Starting pass2
Pass2 complete
Starting pass3
Pass3 complete
Starting pass4
Pass4 complete
Starting pass5
Pass5 complete
Writing changes to disk
fsck.gfs2 complete
Chapter 6. Improving GFS2 performance
There are many aspects of GFS2 configuration that you can analyze to improve file system performance.
For general recommendations for deploying and upgrading Red Hat Enterprise Linux clusters using the High Availability Add-On and Red Hat Global File System 2 (GFS2) see the article Red Hat Enterprise Linux Cluster, High Availability, and GFS Deployment Best Practices on the Red Hat Customer Portal.
6.1. GFS2 file system defragmentation
While there is no defragmentation tool for GFS2 on Red Hat Enterprise Linux, you can defragment individual files by identifying them with the filefrag
tool, copying them to temporary files, and renaming the temporary files to replace the originals.
6.2. GFS2 node locking
In order to get the best performance from a GFS2 file system, it is important to understand some of the basic theory of its operation. A single node file system is implemented alongside a cache, the purpose of which is to eliminate latency of disk accesses when using frequently requested data. In Linux the page cache (and historically the buffer cache) provide this caching function.
With GFS2, each node has its own page cache which may contain some portion of the on-disk data. GFS2 uses a locking mechanism called glocks (pronounced gee-locks) to maintain the integrity of the cache between nodes. The glock subsystem provides a cache management function which is implemented using the distributed lock manager (DLM) as the underlying communication layer.
The glocks provide protection for the cache on a per-inode basis, so there is one lock per inode which is used for controlling the caching layer. If that glock is granted in shared mode (DLM lock mode: PR) then the data under that glock may be cached upon one or more nodes at the same time, so that all the nodes may have local access to the data.
If the glock is granted in exclusive mode (DLM lock mode: EX) then only a single node may cache the data under that glock. This mode is used by all operations which modify the data (such as the write
system call).
If another node requests a glock which cannot be granted immediately, then the DLM sends a message to the node or nodes which currently hold the glocks blocking the new request to ask them to drop their locks. Dropping glocks can be (by the standards of most file system operations) a long process. Dropping a shared glock requires only that the cache be invalidated, which is relatively quick and proportional to the amount of cached data.
Dropping an exclusive glock requires a log flush, and writing back any changed data to disk, followed by the invalidation as per the shared glock.
The difference between a single node file system and GFS2, then, is that a single node file system has a single cache and GFS2 has a separate cache on each node. In both cases, latency to access cached data is of a similar order of magnitude, but the latency to access uncached data is much greater in GFS2 if another node has previously cached that same data.
Operations such as read
(buffered), stat,
and readdir
only require a shared glock. Operations such as write
(buffered), mkdir
, rmdir
, and unlink
require an exclusive glock. Direct I/O read/write operations require a deferred glock if no allocation is taking place, or an exclusive glock if the write requires an allocation (that is, extending the file, or hole filling).
There are two main performance considerations which follow from this. First, read-only operations parallelize extremely well across a cluster, since they can run independently on every node. Second, operations requiring an exclusive glock can reduce performance, if there are multiple nodes contending for access to the same inode(s). Consideration of the working set on each node is thus an important factor in GFS2 file system performance such as when, for example, you perform a file system backup, as described in Backing up a GFS2 file system.
A further consequence of this is that we recommend the use of the noatime
or nodiratime
mount option with GFS2 whenever possible, with the preference for noatime
where the application allows for this. This prevents reads from requiring exclusive locks to update the atime
timestamp.
For users who are concerned about the working set or caching efficiency, GFS2 provides tools that allow you to monitor the performance of a GFS2 file system: Performance Co-Pilot and GFS2 tracepoints.
Due to the way in which GFS2’s caching is implemented the best performance is obtained when either of the following takes place:
- An inode is used in a read-only fashion across all nodes.
- An inode is written or modified from a single node only.
Note that inserting and removing entries from a directory during file creation and deletion counts as writing to the directory inode.
It is possible to break this rule provided that it is broken relatively infrequently. Ignoring this rule too often will result in a severe performance penalty.
If you mmap
() a file on GFS2 with a read/write mapping, but only read from it, this only counts as a read.
If you do not set the noatime
mount
parameter, then reads will also result in writes to update the file timestamps. We recommend that all GFS2 users should mount with noatime
unless they have a specific requirement for atime
.
6.3. Issues with Posix locking
When using Posix locking, you should take the following into account:
- Use of Flocks will yield faster processing than use of Posix locks.
-
Programs using Posix locks in GFS2 should avoid using the
GETLK
function since, in a clustered environment, the process ID may be for a different node in the cluster.
6.4. Performance tuning with GFS2
It is usually possible to alter the way in which a troublesome application stores its data in order to gain a considerable performance advantage.
A typical example of a troublesome application is an email server. These are often laid out with a spool directory containing files for each user (mbox
), or with a directory for each user containing a file for each message (maildir
). When requests arrive over IMAP, the ideal arrangement is to give each user an affinity to a particular node. That way their requests to view and delete email messages will tend to be served from the cache on that one node. Obviously if that node fails, then the session can be restarted on a different node.
When mail arrives by means of SMTP, then again the individual nodes can be set up so as to pass a certain user’s mail to a particular node by default. If the default node is not up, then the message can be saved directly into the user’s mail spool by the receiving node. Again this design is intended to keep particular sets of files cached on just one node in the normal case, but to allow direct access in the case of node failure.
This setup allows the best use of GFS2’s page cache and also makes failures transparent to the application, whether imap
or smtp
.
Backup is often another tricky area. Again, if it is possible it is greatly preferable to back up the working set of each node directly from the node which is caching that particular set of inodes. If you have a backup script which runs at a regular point in time, and that seems to coincide with a spike in the response time of an application running on GFS2, then there is a good chance that the cluster may not be making the most efficient use of the page cache.
Obviously, if you are in the position of being able to stop the application in order to perform a backup, then this will not be a problem. On the other hand, if a backup is run from just one node, then after it has completed a large portion of the file system will be cached on that node, with a performance penalty for subsequent accesses from other nodes. This can be mitigated to a certain extent by dropping the VFS page cache on the backup node after the backup has completed with following command:
echo -n 3 >/proc/sys/vm/drop_caches
However this is not as good a solution as taking care to ensure the working set on each node is either shared, mostly read-only across the cluster, or accessed largely from a single node.
6.5. Troubleshooting GFS2 performance with the GFS2 lock dump
If your cluster performance is suffering because of inefficient use of GFS2 caching, you may see large and increasing I/O wait times. You can make use of GFS2’s lock dump information to determine the cause of the problem.
The GFS2 lock dump information can be gathered from the debugfs
file which can be found at the following path name, assuming that debugfs
is mounted on /sys/kernel/debug/
:
/sys/kernel/debug/gfs2/fsname/glocks
The content of the file is a series of lines. Each line starting with G: represents one glock, and the following lines, indented by a single space, represent an item of information relating to the glock immediately before them in the file.
The best way to use the debugfs
file is to use the cat
command to take a copy of the complete content of the file (it might take a long time if you have a large amount of RAM and a lot of cached inodes) while the application is experiencing problems, and then looking through the resulting data at a later date.
It can be useful to make two copies of the debugfs
file, one a few seconds or even a minute or two after the other. By comparing the holder information in the two traces relating to the same glock number, you can tell whether the workload is making progress (it is just slow) or whether it has become stuck (which is always a bug and should be reported to Red Hat support immediately).
Lines in the debugfs
file starting with H: (holders) represent lock requests either granted or waiting to be granted. The flags field on the holders line f: shows which: The 'W' flag refers to a waiting request, the 'H' flag refers to a granted request. The glocks which have large numbers of waiting requests are likely to be those which are experiencing particular contention.
The following tables show the meanings of the glock flags and glock holder flags.
Flag | Name | Meaning |
---|---|---|
b | Blocking | Valid when the locked flag is set, and indicates that the operation that has been requested from the DLM may block. This flag is cleared for demotion operations and for "try" locks. The purpose of this flag is to allow gathering of stats of the DLM response time independent from the time taken by other nodes to demote locks. |
d | Pending demote | A deferred (remote) demote request |
D | Demote | A demote request (local or remote) |
f | Log flush | The log needs to be committed before releasing this glock |
F | Frozen | Replies from remote nodes ignored - recovery is in progress. This flag is not related to file system freeze, which uses a different mechanism, but is used only in recovery. |
i | Invalidate in progress | In the process of invalidating pages under this glock |
I | Initial | Set when DLM lock is associated with this glock |
l | Locked | The glock is in the process of changing state |
L | LRU | Set when the glock is on the LRU list |
o | Object | Set when the glock is associated with an object (that is, an inode for type 2 glocks, and a resource group for type 3 glocks) |
p | Demote in progress | The glock is in the process of responding to a demote request |
q | Queued | Set when a holder is queued to a glock, and cleared when the glock is held, but there are no remaining holders. Used as part of the algorithm the calculates the minimum hold time for a glock. |
r | Reply pending | Reply received from remote node is awaiting processing |
y | Dirty | Data needs flushing to disk before releasing this glock |
Flag | Name | Meaning |
---|---|---|
a | Async | Do not wait for glock result (will poll for result later) |
A | Any | Any compatible lock mode is acceptable |
c | No cache | When unlocked, demote DLM lock immediately |
e | No expire | Ignore subsequent lock cancel requests |
E | exact | Must have exact lock mode |
F | First | Set when holder is the first to be granted for this lock |
H | Holder | Indicates that requested lock is granted |
p | Priority | Enqueue holder at the head of the queue |
t | Try | A "try" lock |
T | Try 1CB | A "try" lock that sends a callback |
W | Wait | Set while waiting for request to complete |
Having identified a glock which is causing a problem, the next step is to find out which inode it relates to. The glock number (n: on the G: line) indicates this. It is of the form type/number and if type is 2, then the glock is an inode glock and the number is an inode number. To track down the inode, you can then run find -inum number
where number is the inode number converted from the hex format in the glocks file into decimal.
If you run the find
command on a file system when it is experiencing lock contention, you are likely to make the problem worse. It is a good idea to stop the application before running the find
command when you are looking for contended inodes.
The following table shows the meanings of the different glock types.
Type number | Lock type | Use |
---|---|---|
1 | Trans | Transaction lock |
2 | Inode | Inode metadata and data |
3 | Rgrp | Resource group metadata |
4 | Meta | The superblock |
5 | Iopen | Inode last closer detection |
6 | Flock |
|
8 | Quota | Quota operations |
9 | Journal | Journal mutex |
If the glock that was identified was of a different type, then it is most likely to be of type 3: (resource group). If you see significant numbers of processes waiting for other types of glock under normal loads, report this to Red Hat support.
If you do see a number of waiting requests queued on a resource group lock there may be a number of reasons for this. One is that there are a large number of nodes compared to the number of resource groups in the file system. Another is that the file system may be very nearly full (requiring, on average, longer searches for free blocks). The situation in both cases can be improved by adding more storage and using the gfs2_grow
command to expand the file system.
6.6. Enabling data journaling
Ordinarily, GFS2 writes only metadata to its journal. File contents are subsequently written to disk by the kernel’s periodic sync that flushes file system buffers. An fsync()
call on a file causes the file’s data to be written to disk immediately. The call returns when the disk reports that all data is safely written.
Data journaling can result in a reduced fsync()
time for very small files because the file data is written to the journal in addition to the metadata. This advantage rapidly reduces as the file size increases. Writing to medium and larger files will be much slower with data journaling turned on.
Applications that rely on fsync()
to sync file data may see improved performance by using data journaling. Data journaling can be enabled automatically for any GFS2 files created in a flagged directory (and all its subdirectories). Existing files with zero length can also have data journaling turned on or off.
Enabling data journaling on a directory sets the directory to "inherit jdata", which indicates that all files and directories subsequently created in that directory are journaled. You can enable and disable data journaling on a file with the chattr
command.
The following commands enable data journaling on the /mnt/gfs2/gfs2_dir/newfile
file and then check whether the flag has been set properly.
# chattr +j /mnt/gfs2/gfs2_dir/newfile # lsattr /mnt/gfs2/gfs2_dir ---------j--- /mnt/gfs2/gfs2_dir/newfile
The following commands disable data journaling on the /mnt/gfs2/gfs2_dir/newfile
file and then check whether the flag has been set properly.
# chattr -j /mnt/gfs2/gfs2_dir/newfile # lsattr /mnt/gfs2/gfs2_dir ------------- /mnt/gfs2/gfs2_dir/newfile
You can also use the chattr
command to set the j
flag on a directory. When you set this flag for a directory, all files and directories subsequently created in that directory are journaled. The following set of commands sets the j
flag on the gfs2_dir
directory, then checks whether the flag has been set properly. After this, the commands create a new file called newfile
in the /mnt/gfs2/gfs2_dir
directory and then check whether the j
flag has been set for the file. Since the j
flag is set for the directory, then newfile
should also have journaling enabled.
# chattr -j /mnt/gfs2/gfs2_dir # lsattr /mnt/gfs2 ---------j--- /mnt/gfs2/gfs2_dir # touch /mnt/gfs2/gfs2_dir/newfile # lsattr /mnt/gfs2/gfs2_dir ---------j--- /mnt/gfs2/gfs2_dir/newfile
Chapter 7. Diagnosing and correcting problems with GFS2 file systems
The following procedures describe some common GFS2 issues and provide information on how to address them.
7.2. GFS2 file system hangs and requires reboot of one node
If your GFS2 file system hangs and does not return commands run against it, but rebooting one specific node returns the system to normal, this may be indicative of a locking problem or bug. Should this occur, gather GFS2 data during one of these occurences and open a support ticket with Red Hat Support, as described in Gathering GFS2 data for troubleshooting.
7.3. GFS2 file system hangs and requires reboot of all nodes
If your GFS2 file system hangs and does not return commands run against it, requiring that you reboot all nodes in the cluster before using it, check for the following issues.
- You may have had a failed fence. GFS2 file systems will freeze to ensure data integrity in the event of a failed fence. Check the messages logs to see if there are any failed fences at the time of the hang. Ensure that fencing is configured correctly.
The GFS2 file system may have withdrawn. Check through the messages logs for the word
withdraw
and check for any messages and call traces from GFS2 indicating that the file system has been withdrawn. A withdraw is indicative of file system corruption, a storage failure, or a bug. At the earliest time when it is convenient to unmount the file system, you should perform the following procedure:Reboot the node on which the withdraw occurred.
# /sbin/reboot
Stop the file system resource to unmount the GFS2 file system on all nodes.
# pcs resource disable --wait=100 mydata_fs
Capture the metadata with the
gfs2_edit savemeta…
command. You should ensure that there is sufficient space for the file, which in some cases may be large. In this example, the metadata is saved to a file in the/root
directory.# gfs2_edit savemeta /dev/vg_mydata/mydata /root/gfs2metadata.gz
Update the
gfs2-utils
package.# sudo dnf update gfs2-utils
On one node, run the
fsck.gfs2
command on the file system to ensure file system integrity and repair any damage.# fsck.gfs2 -y /dev/vg_mydata/mydata > /tmp/fsck.out
After the
fsck.gfs2
command has completed, re-enable the file system resource to return it to service:# pcs resource enable --wait=100 mydata_fs
Open a support ticket with Red Hat Support. Inform them you experienced a GFS2 withdraw and provide logs and the debugging information generated by the
sosreports
andgfs2_edit savemeta
commands.In some instances of a GFS2 withdraw, commands can hang that are trying to access the file system or its block device. In these cases a hard reboot is required to reboot the cluster.
For information about the GFS2 withdraw function, see GFS2 filesystem unavailable to a node (the GFS2 withdraw function).
- This error may be indicative of a locking problem or bug. Gather data during one of these occurrences and open a support ticket with Red Hat Support, as described in Gathering GFS2 data for troubleshooting.
7.4. GFS2 file system does not mount on newly added cluster node
If you add a new node to a cluster and find that you cannot mount your GFS2 file system on that node, you may have fewer journals on the GFS2 file system than nodes attempting to access the GFS2 file system. You must have one journal per GFS2 host you intend to mount the file system on (with the exception of GFS2 file systems mounted with the spectator
mount option set, since these do not require a journal). You can add journals to a GFS2 file system with the gfs2_jadd
command, as described in Adding journals to a GFS2 file system.
7.5. Space indicated as used in empty file system
If you have an empty GFS2 file system, the df
command will show that there is space being taken up. This is because GFS2 file system journals consume space (number of journals * journal size) on disk. f you created a GFS2 file system with a large number of journals or specified a large journal size then you will be see (number of journals * journal size) as already in use when you execute the df
command. Even if you did not specify a large number of journals or large journals, small GFS2 file systems (in the 1GB or less range) will show a large amount of space as being in use with the default GFS2 journal size.
7.6. Gathering GFS2 data for troubleshooting
If your GFS2 file system hangs and does not return commands run against it and you find that you need to open a ticket with Red Hat Support, you should first gather the following data:
The GFS2 lock dump for the file system on each node:
cat /sys/kernel/debug/gfs2/fsname/glocks >glocks.fsname.nodename
The DLM lock dump for the file system on each node: You can get this information with the
dlm_tool
:dlm_tool lockdebug -sv lsname
In this command, lsname is the lockspace name used by DLM for the file system in question. You can find this value in the output from the
group_tool
command.-
The output from the
sysrq -t
command. -
The contents of the
/var/log/messages
file.
Once you have gathered that data, you can open a ticket with Red Hat Support and provide the data you have collected.
Chapter 8. GFS2 file systems in a cluster
Use the following administrative procedures to configure GFS2 file systems in a Red Hat high availability cluster.
8.1. Configuring a GFS2 file system in a cluster
You can set up a Pacemaker cluster that includes GFS2 file systems with the following procedure. In this example, you create three GFS2 file systems on three logical volumes in a two-node cluster.
Prerequisites
- Install and start the cluster software on both cluster nodes and create a basic two-node cluster.
- Configure fencing for the cluster.
For information about creating a Pacemaker cluster and configuring fencing for the cluster, see Creating a Red Hat High-Availability cluster with Pacemaker.
Procedure
On both nodes in the cluster, enable the repository for Resilient Storage that corresponds to your system architecture. For example, to enable the Resilient Storage repository for an x86_64 system, you can enter the following
subscription-manager
command:# subscription-manager repos --enable=rhel-9-for-x86_64-resilientstorage-rpms
Note that the Resilient Storage repository is a superset of the High Availability repository. If you enable the Resilient Storage repository you do not also need to enable the High Availability repository.
On both nodes of the cluster, install the
lvm2-lockd
,gfs2-utils
, anddlm
packages. To support these packages, you must be subscribed to the AppStream channel and the Resilient Storage channel.# dnf install lvm2-lockd gfs2-utils dlm
On both nodes of the cluster, set the
use_lvmlockd
configuration option in the/etc/lvm/lvm.conf
file touse_lvmlockd=1
.... use_lvmlockd = 1 ...
Set the global Pacemaker parameter
no-quorum-policy
tofreeze
.NoteBy default, the value of
no-quorum-policy
is set tostop
, indicating that once quorum is lost, all the resources on the remaining partition will immediately be stopped. Typically this default is the safest and most optimal option, but unlike most resources, GFS2 requires quorum to function. When quorum is lost both the applications using the GFS2 mounts and the GFS2 mount itself cannot be correctly stopped. Any attempts to stop these resources without quorum will fail which will ultimately result in the entire cluster being fenced every time quorum is lost.To address this situation, set
no-quorum-policy
tofreeze
when GFS2 is in use. This means that when quorum is lost, the remaining partition will do nothing until quorum is regained.[root@z1 ~]# pcs property set no-quorum-policy=freeze
Set up a
dlm
resource. This is a required dependency for configuring a GFS2 file system in a cluster. This example creates thedlm
resource as part of a resource group namedlocking
.[root@z1 ~]# pcs resource create dlm --group locking ocf:pacemaker:controld op monitor interval=30s on-fail=fence
Clone the
locking
resource group so that the resource group can be active on both nodes of the cluster.[root@z1 ~]# pcs resource clone locking interleave=true
Set up an
lvmlockd
resource as part of thelocking
resource group.[root@z1 ~]# pcs resource create lvmlockd --group locking ocf:heartbeat:lvmlockd op monitor interval=30s on-fail=fence
Check the status of the cluster to ensure that the
locking
resource group has started on both nodes of the cluster.[root@z1 ~]# pcs status --full Cluster name: my_cluster [...] Online: [ z1.example.com (1) z2.example.com (2) ] Full list of resources: smoke-apc (stonith:fence_apc): Started z1.example.com Clone Set: locking-clone [locking] Resource Group: locking:0 dlm (ocf::pacemaker:controld): Started z1.example.com lvmlockd (ocf::heartbeat:lvmlockd): Started z1.example.com Resource Group: locking:1 dlm (ocf::pacemaker:controld): Started z2.example.com lvmlockd (ocf::heartbeat:lvmlockd): Started z2.example.com Started: [ z1.example.com z2.example.com ]
On one node of the cluster, create two shared volume groups. One volume group will contain two GFS2 file systems, and the other volume group will contain one GFS2 file system.
NoteIf your LVM volume group contains one or more physical volumes that reside on remote block storage, such as an iSCSI target, Red Hat recommends that you ensure that the service starts before Pacemaker starts. For information about configuring startup order for a remote physical volume used by a Pacemaker cluster, see Configuring startup order for resource dependencies not managed by Pacemaker.
The following command creates the shared volume group
shared_vg1
on/dev/vdb
.[root@z1 ~]# vgcreate --shared shared_vg1 /dev/vdb Physical volume "/dev/vdb" successfully created. Volume group "shared_vg1" successfully created VG shared_vg1 starting dlm lockspace Starting locking. Waiting until locks are ready...
The following command creates the shared volume group
shared_vg2
on/dev/vdc
.[root@z1 ~]# vgcreate --shared shared_vg2 /dev/vdc Physical volume "/dev/vdc" successfully created. Volume group "shared_vg2" successfully created VG shared_vg2 starting dlm lockspace Starting locking. Waiting until locks are ready...
On the second node in the cluster:
If the use of a devices file is enabled with the
use_devicesfile = 1
parameter in thelvm.conf
file, add the shared devices to the devices file This feature is enabled by default.[root@z2 ~]# lvmdevices --adddev /dev/vdb [root@z2 ~]# lvmdevices --adddev /dev/vdc
Start the lock manager for each of the shared volume groups.
[root@z2 ~]# vgchange --lockstart shared_vg1 VG shared_vg1 starting dlm lockspace Starting locking. Waiting until locks are ready... [root@z2 ~]# vgchange --lockstart shared_vg2 VG shared_vg2 starting dlm lockspace Starting locking. Waiting until locks are ready...
On one node in the cluster, create the shared logical volumes and format the volumes with a GFS2 file system. One journal is required for each node that mounts the file system. Ensure that you create enough journals for each of the nodes in your cluster. The format of the lock table name is ClusterName:FSName where ClusterName is the name of the cluster for which the GFS2 file system is being created and FSName is the file system name, which must be unique for all
lock_dlm
file systems over the cluster.[root@z1 ~]# lvcreate --activate sy -L5G -n shared_lv1 shared_vg1 Logical volume "shared_lv1" created. [root@z1 ~]# lvcreate --activate sy -L5G -n shared_lv2 shared_vg1 Logical volume "shared_lv2" created. [root@z1 ~]# lvcreate --activate sy -L5G -n shared_lv1 shared_vg2 Logical volume "shared_lv1" created. [root@z1 ~]# mkfs.gfs2 -j2 -p lock_dlm -t my_cluster:gfs2-demo1 /dev/shared_vg1/shared_lv1 [root@z1 ~]# mkfs.gfs2 -j2 -p lock_dlm -t my_cluster:gfs2-demo2 /dev/shared_vg1/shared_lv2 [root@z1 ~]# mkfs.gfs2 -j2 -p lock_dlm -t my_cluster:gfs2-demo3 /dev/shared_vg2/shared_lv1
Create an
LVM-activate
resource for each logical volume to automatically activate that logical volume on all nodes.Create an
LVM-activate
resource namedsharedlv1
for the logical volumeshared_lv1
in volume groupshared_vg1
. This command also creates the resource groupshared_vg1
that includes the resource. In this example, the resource group has the same name as the shared volume group that includes the logical volume.[root@z1 ~]# pcs resource create sharedlv1 --group shared_vg1 ocf:heartbeat:LVM-activate lvname=shared_lv1 vgname=shared_vg1 activation_mode=shared vg_access_mode=lvmlockd
Create an
LVM-activate
resource namedsharedlv2
for the logical volumeshared_lv2
in volume groupshared_vg1
. This resource will also be part of the resource groupshared_vg1
.[root@z1 ~]# pcs resource create sharedlv2 --group shared_vg1 ocf:heartbeat:LVM-activate lvname=shared_lv2 vgname=shared_vg1 activation_mode=shared vg_access_mode=lvmlockd
Create an
LVM-activate
resource namedsharedlv3
for the logical volumeshared_lv1
in volume groupshared_vg2
. This command also creates the resource groupshared_vg2
that includes the resource.[root@z1 ~]# pcs resource create sharedlv3 --group shared_vg2 ocf:heartbeat:LVM-activate lvname=shared_lv1 vgname=shared_vg2 activation_mode=shared vg_access_mode=lvmlockd
Clone the two new resource groups.
[root@z1 ~]# pcs resource clone shared_vg1 interleave=true [root@z1 ~]# pcs resource clone shared_vg2 interleave=true
Configure ordering constraints to ensure that the
locking
resource group that includes thedlm
andlvmlockd
resources starts first.[root@z1 ~]# pcs constraint order start locking-clone then shared_vg1-clone Adding locking-clone shared_vg1-clone (kind: Mandatory) (Options: first-action=start then-action=start) [root@z1 ~]# pcs constraint order start locking-clone then shared_vg2-clone Adding locking-clone shared_vg2-clone (kind: Mandatory) (Options: first-action=start then-action=start)
Configure colocation constraints to ensure that the
vg1
andvg2
resource groups start on the same node as thelocking
resource group.[root@z1 ~]# pcs constraint colocation add shared_vg1-clone with locking-clone [root@z1 ~]# pcs constraint colocation add shared_vg2-clone with locking-clone
On both nodes in the cluster, verify that the logical volumes are active. There may be a delay of a few seconds.
[root@z1 ~]# lvs LV VG Attr LSize shared_lv1 shared_vg1 -wi-a----- 5.00g shared_lv2 shared_vg1 -wi-a----- 5.00g shared_lv1 shared_vg2 -wi-a----- 5.00g [root@z2 ~]# lvs LV VG Attr LSize shared_lv1 shared_vg1 -wi-a----- 5.00g shared_lv2 shared_vg1 -wi-a----- 5.00g shared_lv1 shared_vg2 -wi-a----- 5.00g
Create a file system resource to automatically mount each GFS2 file system on all nodes.
You should not add the file system to the
/etc/fstab
file because it will be managed as a Pacemaker cluster resource. Mount options can be specified as part of the resource configuration withoptions=options
. Run thepcs resource describe Filesystem
command to display the full configuration options.The following commands create the file system resources. These commands add each resource to the resource group that includes the logical volume resource for that file system.
[root@z1 ~]# pcs resource create sharedfs1 --group shared_vg1 ocf:heartbeat:Filesystem device="/dev/shared_vg1/shared_lv1" directory="/mnt/gfs1" fstype="gfs2" options=noatime op monitor interval=10s on-fail=fence [root@z1 ~]# pcs resource create sharedfs2 --group shared_vg1 ocf:heartbeat:Filesystem device="/dev/shared_vg1/shared_lv2" directory="/mnt/gfs2" fstype="gfs2" options=noatime op monitor interval=10s on-fail=fence [root@z1 ~]# pcs resource create sharedfs3 --group shared_vg2 ocf:heartbeat:Filesystem device="/dev/shared_vg2/shared_lv1" directory="/mnt/gfs3" fstype="gfs2" options=noatime op monitor interval=10s on-fail=fence
Verification
Verify that the GFS2 file systems are mounted on both nodes of the cluster.
[root@z1 ~]# mount | grep gfs2 /dev/mapper/shared_vg1-shared_lv1 on /mnt/gfs1 type gfs2 (rw,noatime,seclabel) /dev/mapper/shared_vg1-shared_lv2 on /mnt/gfs2 type gfs2 (rw,noatime,seclabel) /dev/mapper/shared_vg2-shared_lv1 on /mnt/gfs3 type gfs2 (rw,noatime,seclabel) [root@z2 ~]# mount | grep gfs2 /dev/mapper/shared_vg1-shared_lv1 on /mnt/gfs1 type gfs2 (rw,noatime,seclabel) /dev/mapper/shared_vg1-shared_lv2 on /mnt/gfs2 type gfs2 (rw,noatime,seclabel) /dev/mapper/shared_vg2-shared_lv1 on /mnt/gfs3 type gfs2 (rw,noatime,seclabel)
Check the status of the cluster.
[root@z1 ~]# pcs status --full Cluster name: my_cluster [...] Full list of resources: smoke-apc (stonith:fence_apc): Started z1.example.com Clone Set: locking-clone [locking] Resource Group: locking:0 dlm (ocf::pacemaker:controld): Started z2.example.com lvmlockd (ocf::heartbeat:lvmlockd): Started z2.example.com Resource Group: locking:1 dlm (ocf::pacemaker:controld): Started z1.example.com lvmlockd (ocf::heartbeat:lvmlockd): Started z1.example.com Started: [ z1.example.com z2.example.com ] Clone Set: shared_vg1-clone [shared_vg1] Resource Group: shared_vg1:0 sharedlv1 (ocf::heartbeat:LVM-activate): Started z2.example.com sharedlv2 (ocf::heartbeat:LVM-activate): Started z2.example.com sharedfs1 (ocf::heartbeat:Filesystem): Started z2.example.com sharedfs2 (ocf::heartbeat:Filesystem): Started z2.example.com Resource Group: shared_vg1:1 sharedlv1 (ocf::heartbeat:LVM-activate): Started z1.example.com sharedlv2 (ocf::heartbeat:LVM-activate): Started z1.example.com sharedfs1 (ocf::heartbeat:Filesystem): Started z1.example.com sharedfs2 (ocf::heartbeat:Filesystem): Started z1.example.com Started: [ z1.example.com z2.example.com ] Clone Set: shared_vg2-clone [shared_vg2] Resource Group: shared_vg2:0 sharedlv3 (ocf::heartbeat:LVM-activate): Started z2.example.com sharedfs3 (ocf::heartbeat:Filesystem): Started z2.example.com Resource Group: shared_vg2:1 sharedlv3 (ocf::heartbeat:LVM-activate): Started z1.example.com sharedfs3 (ocf::heartbeat:Filesystem): Started z1.example.com Started: [ z1.example.com z2.example.com ] ...
8.2. Configuring an encrypted GFS2 file system in a cluster
You can create a Pacemaker cluster that includes a LUKS encrypted GFS2 file system with the following procedure. In this example, you create one GFS2 file systems on a logical volume and encrypt the file system. Encrypted GFS2 file systems are supported using the crypt
resource agent, which provides support for LUKS encryption.
There are three parts to this procedure:
- Configuring a shared logical volume in a Pacemaker cluster
-
Encrypting the logical volume and creating a
crypt
resource - Formatting the encrypted logical volume with a GFS2 file system and creating a file system resource for the cluster
8.2.2. Encrypt the logical volume and create a crypt resource
Prerequisites
- You have configured a shared logical volume in a Pacemaker cluster.
Procedure
On one node in the cluster, create a new file that will contain the crypt key and set the permissions on the file so that it is readable only by root.
[root@z1 ~]# touch /etc/crypt_keyfile [root@z1 ~]# chmod 600 /etc/crypt_keyfile
Create the crypt key.
[root@z1 ~]# dd if=/dev/urandom bs=4K count=1 of=/etc/crypt_keyfile 1+0 records in 1+0 records out 4096 bytes (4.1 kB, 4.0 KiB) copied, 0.000306202 s, 13.4 MB/s [root@z1 ~]# scp /etc/crypt_keyfile root@z2.example.com:/etc/
Distribute the crypt keyfile to the other nodes in the cluster, using the
-p
parameter to preserve the permissions you set.[root@z1 ~]# scp -p /etc/crypt_keyfile root@z2.example.com:/etc/
Create the encrypted device on the LVM volume where you will configure the encrypted GFS2 file system.
[root@z1 ~]# cryptsetup luksFormat /dev/shared_vg1/shared_lv1 --type luks2 --key-file=/etc/crypt_keyfile WARNING! ======== This will overwrite data on /dev/shared_vg1/shared_lv1 irrevocably. Are you sure? (Type 'yes' in capital letters): YES
Create the crypt resource as part of the
shared_vg1
volume group.[root@z1 ~]# pcs resource create crypt --group shared_vg1 ocf:heartbeat:crypt crypt_dev="luks_lv1" crypt_type=luks2 key_file=/etc/crypt_keyfile encrypted_dev="/dev/shared_vg1/shared_lv1"
Verification
Ensure that the crypt resource has created the crypt device, which in this example is /dev/mapper/luks_lv1
.
[root@z1 ~]# ls -l /dev/mapper/
...
lrwxrwxrwx 1 root root 7 Mar 4 09:52 luks_lv1 -> ../dm-3
...
8.2.3. Format the encrypted logical volume with a GFS2 file system and create a file system resource for the cluster
Prerequisites
- You have encrypted the logical volume and created a crypt resource.
Procedure
On one node in the cluster, format the volume with a GFS2 file system. One journal is required for each node that mounts the file system. Ensure that you create enough journals for each of the nodes in your cluster. The format of the lock table name is ClusterName:FSName where ClusterName is the name of the cluster for which the GFS2 file system is being created and FSName is the file system name, which must be unique for all
lock_dlm
file systems over the cluster.[root@z1 ~]# mkfs.gfs2 -j3 -p lock_dlm -t my_cluster:gfs2-demo1 /dev/mapper/luks_lv1 /dev/mapper/luks_lv1 is a symbolic link to /dev/dm-3 This will destroy any data on /dev/dm-3 Are you sure you want to proceed? [y/n] y Discarding device contents (may take a while on large devices): Done Adding journals: Done Building resource groups: Done Creating quota file: Done Writing superblock and syncing: Done Device: /dev/mapper/luks_lv1 Block size: 4096 Device size: 4.98 GB (1306624 blocks) Filesystem size: 4.98 GB (1306622 blocks) Journals: 3 Journal size: 16MB Resource groups: 23 Locking protocol: "lock_dlm" Lock table: "my_cluster:gfs2-demo1" UUID: de263f7b-0f12-4d02-bbb2-56642fade293
Create a file system resource to automatically mount the GFS2 file system on all nodes.
Do not add the file system to the
/etc/fstab
file because it will be managed as a Pacemaker cluster resource. Mount options can be specified as part of the resource configuration withoptions=options
. Run thepcs resource describe Filesystem
command for full configuration options.The following command creates the file system resource. This command adds the resource to the resource group that includes the logical volume resource for that file system.
[root@z1 ~]# pcs resource create sharedfs1 --group shared_vg1 ocf:heartbeat:Filesystem device="/dev/mapper/luks_lv1" directory="/mnt/gfs1" fstype="gfs2" options=noatime op monitor interval=10s on-fail=fence
Verification
Verify that the GFS2 file system is mounted on both nodes of the cluster.
[root@z1 ~]# mount | grep gfs2 /dev/mapper/luks_lv1 on /mnt/gfs1 type gfs2 (rw,noatime,seclabel) [root@z2 ~]# mount | grep gfs2 /dev/mapper/luks_lv1 on /mnt/gfs1 type gfs2 (rw,noatime,seclabel)
Check the status of the cluster.
[root@z1 ~]# pcs status --full Cluster name: my_cluster [...] Full list of resources: smoke-apc (stonith:fence_apc): Started z1.example.com Clone Set: locking-clone [locking] Resource Group: locking:0 dlm (ocf::pacemaker:controld): Started z2.example.com lvmlockd (ocf::heartbeat:lvmlockd): Started z2.example.com Resource Group: locking:1 dlm (ocf::pacemaker:controld): Started z1.example.com lvmlockd (ocf::heartbeat:lvmlockd): Started z1.example.com Started: [ z1.example.com z2.example.com ] Clone Set: shared_vg1-clone [shared_vg1] Resource Group: shared_vg1:0 sharedlv1 (ocf::heartbeat:LVM-activate): Started z2.example.com crypt (ocf::heartbeat:crypt) Started z2.example.com sharedfs1 (ocf::heartbeat:Filesystem): Started z2.example.com Resource Group: shared_vg1:1 sharedlv1 (ocf::heartbeat:LVM-activate): Started z1.example.com crypt (ocf::heartbeat:crypt) Started z1.example.com sharedfs1 (ocf::heartbeat:Filesystem): Started z1.example.com Started: [z1.example.com z2.example.com ] ...
Additional resources
Chapter 9. GFS2 tracepoints and the glock debugfs interface
This documentation on both the GFS2 tracepoints and the glock debugfs
interface is intended for advanced users who are familiar with file system internals and who would like to learn more about the design of GFS2 and how to debug GFS2-specific issues.
The following sections describe GFS2 tracepoints and the GFS2 glocks
file.
9.1. GFS2 tracepoint types
There are currently three types of GFS2 tracepoints: glock (pronounced "gee-lock") tracepoints, bmap tracepoints and log tracepoints. These can be used to monitor a running GFS2 file system. Tracepoints are particularly useful when a problem, such as a hang or performance issue, is reproducible and thus the tracepoint output can be obtained during the problematic operation. In GFS2, glocks are the primary cache control mechanism and they are the key to understanding the performance of the core of GFS2. The bmap (block map) tracepoints can be used to monitor block allocations and block mapping (lookup of already allocated blocks in the on-disk metadata tree) as they happen and check for any issues relating to locality of access. The log tracepoints keep track of the data being written to and released from the journal and can provide useful information about that part of GFS2.
The tracepoints are designed to be as generic as possible. This should mean that it will not be necessary to change the API during the course of Red Hat Enterprise Linux 9. On the other hand, users of this interface should be aware that this is a debugging interface and not part of the normal Red Hat Enterprise Linux 9 API set, and as such Red Hat makes no guarantees that changes in the GFS2 tracepoints interface will not occur.
Tracepoints are a generic feature of Red Hat Enterprise Linux and their scope goes well beyond GFS2. In particular they are used to implement the blktrace
infrastructure and the blktrace
tracepoints can be used in combination with those of GFS2 to gain a fuller picture of the system performance. Due to the level at which the tracepoints operate, they can produce large volumes of data in a very short period of time. They are designed to put a minimum load on the system when they are enabled, but it is inevitable that they will have some effect. Filtering events by a variety of means can help reduce the volume of data and help focus on obtaining just the information which is useful for understanding any particular situation.
9.2. Tracepoints
The tracepoints can be found under the /sys/kernel/debug/tracing/
directory assuming that debugfs
is mounted in the standard place at the /sys/kernel/debug
directory. The events
subdirectory contains all the tracing events that may be specified and, provided the gfs2
module is loaded, there will be a gfs2
subdirectory containing further subdirectories, one for each GFS2 event. The contents of the /sys/kernel/debug/tracing/events/gfs2
directory should look roughly like the following:
[root@chywoon gfs2]# ls
enable gfs2_bmap gfs2_glock_queue gfs2_log_flush
filter gfs2_demote_rq gfs2_glock_state_change gfs2_pin
gfs2_block_alloc gfs2_glock_put gfs2_log_blocks gfs2_promote
To enable all the GFS2 tracepoints, enter the following command:
[root@chywoon gfs2]# echo -n 1 >/sys/kernel/debug/tracing/events/gfs2/enable
To enable a specific tracepoint, there is an enable
file in each of the individual event subdirectories. The same is true of the filter
file which can be used to set an event filter for each event or set of events. The meaning of the individual events is explained in more detail below.
The output from the tracepoints is available in ASCII or binary format. This appendix does not currently cover the binary interface. The ASCII interface is available in two ways. To list the current content of the ring buffer, you can enter the following command:
[root@chywoon gfs2]# cat /sys/kernel/debug/tracing/trace
This interface is useful in cases where you are using a long-running process for a certain period of time and, after some event, want to look back at the latest captured information in the buffer. An alternative interface, /sys/kernel/debug/tracing/trace_pipe
, can be used when all the output is required. Events are read from this file as they occur; there is no historical information available through this interface. The format of the output is the same from both interfaces and is described for each of the GFS2 events in the later sections of this appendix.
A utility called trace-cmd
is available for reading tracepoint data. For more information about this utility, see http://lwn.net/Articles/341902/. The trace-cmd
utility can be used in a similar way to the strace
utility, for example to run a command while gathering trace data from various sources.
9.3. Glocks
To understand GFS2, the most important concept to understand, and the one which sets it aside from other file systems, is the concept of glocks. In terms of the source code, a glock is a data structure that brings together the DLM and caching into a single state machine. Each glock has a 1:1 relationship with a single DLM lock, and provides caching for that lock state so that repetitive operations carried out from a single node of the file system do not have to repeatedly call the DLM, and thus they help avoid unnecessary network traffic. There are two broad categories of glocks, those which cache metadata and those which do not. The inode glocks and the resource group glocks both cache metadata, other types of glocks do not cache metadata. The inode glock is also involved in the caching of data in addition to metadata and has the most complex logic of all glocks.
Glock mode | DLM lock mode | Notes |
---|---|---|
UN | IV/NL | Unlocked (no DLM lock associated with glock or NL lock depending on I flag) |
SH | PR | Shared (protected read) lock |
EX | EX | Exclusive lock |
DF | CW | Deferred (concurrent write) used for Direct I/O and file system freeze |
Glocks remain in memory until either they are unlocked (at the request of another node or at the request of the VM) and there are no local users. At that point they are removed from the glock hash table and freed. When a glock is created, the DLM lock is not associated with the glock immediately. The DLM lock becomes associated with the glock upon the first request to the DLM, and if this request is successful then the 'I' (initial) flag will be set on the glock. The "Glock Flags" table in The glock debugfs interface shows the meanings of the different glock flags. Once the DLM has been associated with the glock, the DLM lock will always remain at least at NL (Null) lock mode until the glock is to be freed. A demotion of the DLM lock from NL to unlocked is always the last operation in the life of a glock.
Each glock can have a number of "holders" associated with it, each of which represents one lock request from the higher layers. System calls relating to GFS2 queue and dequeue holders from the glock to protect the critical section of code.
The glock state machine is based on a work queue. For performance reasons, tasklets would be preferable; however, in the current implementation we need to submit I/O from that context which prohibits their use.
Workqueues have their own tracepoints which can be used in combination with the GFS2 tracepoints.
The following table shows what state may be cached under each of the glock modes and whether that cached state may be dirty. This applies to both inode and resource group locks, although there is no data component for the resource group locks, only metadata.
Glock mode | Cache Data | Cache Metadata | Dirty Data | Dirty Metadata |
---|---|---|---|---|
UN | No | No | No | No |
SH | Yes | Yes | No | No |
DF | No | Yes | No | No |
EX | Yes | Yes | Yes | Yes |
9.4. The glock debugfs interface
The glock debugfs
interface allows the visualization of the internal state of the glocks and the holders and it also includes some summary details of the objects being locked in some cases. Each line of the file either begins G: with no indentation (which refers to the glock itself) or it begins with a different letter, indented with a single space, and refers to the structures associated with the glock immediately above it in the file (H: is a holder, I: an inode, and R: a resource group). Here is an example of what the content of this file might look like:
G: s:SH n:5/75320 f:I t:SH d:EX/0 a:0 r:3 H: s:SH f:EH e:0 p:4466 [postmark] gfs2_inode_lookup+0x14e/0x260 [gfs2] G: s:EX n:3/258028 f:yI t:EX d:EX/0 a:3 r:4 H: s:EX f:tH e:0 p:4466 [postmark] gfs2_inplace_reserve_i+0x177/0x780 [gfs2] R: n:258028 f:05 b:22256/22256 i:16800 G: s:EX n:2/219916 f:yfI t:EX d:EX/0 a:0 r:3 I: n:75661/219916 t:8 f:0x10 d:0x00000000 s:7522/7522 G: s:SH n:5/127205 f:I t:SH d:EX/0 a:0 r:3 H: s:SH f:EH e:0 p:4466 [postmark] gfs2_inode_lookup+0x14e/0x260 [gfs2] G: s:EX n:2/50382 f:yfI t:EX d:EX/0 a:0 r:2 G: s:SH n:5/302519 f:I t:SH d:EX/0 a:0 r:3 H: s:SH f:EH e:0 p:4466 [postmark] gfs2_inode_lookup+0x14e/0x260 [gfs2] G: s:SH n:5/313874 f:I t:SH d:EX/0 a:0 r:3 H: s:SH f:EH e:0 p:4466 [postmark] gfs2_inode_lookup+0x14e/0x260 [gfs2] G: s:SH n:5/271916 f:I t:SH d:EX/0 a:0 r:3 H: s:SH f:EH e:0 p:4466 [postmark] gfs2_inode_lookup+0x14e/0x260 [gfs2] G: s:SH n:5/312732 f:I t:SH d:EX/0 a:0 r:3 H: s:SH f:EH e:0 p:4466 [postmark] gfs2_inode_lookup+0x14e/0x260 [gfs2]
The above example is a series of excerpts (from an approximately 18MB file) generated by the command cat /sys/kernel/debug/gfs2/unity:myfs/glocks >my.lock
during a run of the postmark benchmark on a single node GFS2 file system. The glocks in the figure have been selected in order to show some of the more interesting features of the glock dumps.
The glock states are either EX (exclusive), DF (deferred), SH (shared) or UN (unlocked). These states correspond directly with DLM lock modes except for UN which may represent either the DLM null lock state, or that GFS2 does not hold a DLM lock (depending on the I flag as explained above). The s: field of the glock indicates the current state of the lock and the same field in the holder indicates the requested mode. If the lock is granted, the holder will have the H bit set in its flags (f: field). Otherwise, it will have the W wait bit set.
The n: field (number) indicates the number associated with each item. For glocks, that is the type number followed by the glock number so that in the above example, the first glock is n:5/75320; which indicates an iopen
glock which relates to inode 75320. In the case of inode and iopen
glocks, the glock number is always identical to the inode’s disk block number.
The glock numbers (n: field) in the debugfs glocks file are in hexadecimal, whereas the tracepoints output lists them in decimal. This is for historical reasons; glock numbers were always written in hex, but decimal was chosen for the tracepoints so that the numbers could easily be compared with the other tracepoint output (from blktrace
for example) and with output from stat
(1).
The full listing of all the flags for both the holder and the glock are set out in the "Glock Flags" table, below, and the "Glock Holder Flags" table in Glock holders. The content of lock value blocks is not currently available through the glock debugfs
interface. The following table shows the meanings of the different glock types.
Type number | Lock type | Use |
---|---|---|
1 | trans | Transaction lock |
2 | inode | Inode metadata and data |
3 | rgrp | Resource group metadata |
4 | meta | The superblock |
5 | iopen | Inode last closer detection |
6 | flock |
|
8 | quota | Quota operations |
9 | journal | Journal mutex |
One of the more important glock flags is the l (locked) flag. This is the bit lock that is used to arbitrate access to the glock state when a state change is to be performed. It is set when the state machine is about to send a remote lock request through the DLM, and only cleared when the complete operation has been performed. Sometimes this can mean that more than one lock request will have been sent, with various invalidations occurring between times.
The following table shows the meanings of the different glock flags.
Flag | Name | Meaning |
---|---|---|
d | Pending demote | A deferred (remote) demote request |
D | Demote | A demote request (local or remote) |
f | Log flush | The log needs to be committed before releasing this glock |
F | Frozen | Replies from remote nodes ignored - recovery is in progress. |
i | Invalidate in progress | In the process of invalidating pages under this glock |
I | Initial | Set when DLM lock is associated with this glock |
l | Locked | The glock is in the process of changing state |
L | LRU | Set when the glock is on the LRU list` |
o | Object | Set when the glock is associated with an object (that is, an inode for type 2 glocks, and a resource group for type 3 glocks) |
p | Demote in progress | The glock is in the process of responding to a demote request |
q | Queued | Set when a holder is queued to a glock, and cleared when the glock is held, but there are no remaining holders. Used as part of the algorithm the calculates the minimum hold time for a glock. |
r | Reply pending | Reply received from remote node is awaiting processing |
y | Dirty | Data needs flushing to disk before releasing this glock |
When a remote callback is received from a node that wants to get a lock in a mode that conflicts with that being held on the local node, then one or other of the two flags D (demote) or d (demote pending) is set. In order to prevent starvation conditions when there is contention on a particular lock, each lock is assigned a minimum hold time. A node which has not yet had the lock for the minimum hold time is allowed to retain that lock until the time interval has expired.
If the time interval has expired, then the D (demote) flag will be set and the state required will be recorded. In that case the next time there are no granted locks on the holders queue, the lock will be demoted. If the time interval has not expired, then the d (demote pending) flag is set instead. This also schedules the state machine to clear d (demote pending) and set D (demote) when the minimum hold time has expired.
The I (initial) flag is set when the glock has been assigned a DLM lock. This happens when the glock is first used and the I flag will then remain set until the glock is finally freed (which the DLM lock is unlocked).
9.5. Glock holders
The following table shows the meanings of the different glock holder flags.
Flag | Name | Meaning |
---|---|---|
a | Async | Do not wait for glock result (will poll for result later) |
A | Any | Any compatible lock mode is acceptable |
c | No cache | When unlocked, demote DLM lock immediately |
e | No expire | Ignore subsequent lock cancel requests |
E | Exact | Must have exact lock mode |
F | First | Set when holder is the first to be granted for this lock |
H | Holder | Indicates that requested lock is granted |
p | Priority | Enqueue holder at the head of the queue |
t | Try | A "try" lock |
T | Try 1CB | A "try" lock that sends a callback |
W | Wait | Set while waiting for request to complete |
The most important holder flags are H (holder) and W (wait) as mentioned earlier, since they are set on granted lock requests and queued lock requests respectively. The ordering of the holders in the list is important. If there are any granted holders, they will always be at the head of the queue, followed by any queued holders.
If there are no granted holders, then the first holder in the list will be the one that triggers the next state change. Since demote requests are always onsidered higher priority than requests from the file system, that might not always directly result in a change to the state requested.
The glock subsystem supports two kinds of "try" lock. These are useful both because they allow the taking of locks out of the normal order (with suitable back-off and retry) and because they can be used to help avoid resources in use by other nodes. The normal t (try) lock is just what its name indicates; it is a "try" lock that does not do anything special. The T (try 1CB
) lock, on the other hand, is identical to the t lock except that the DLM will send a single callback to current incompatible lock holders. One use of the T (try 1CB
) lock is with the iopen
locks, which are used to arbitrate among the nodes when an inode’s i_nlink
count is zero, and determine which of the nodes will be responsible for deallocating the inode. The iopen
glock is normally held in the shared state, but when the i_nlink
count becomes zero and →evict_inode
() is called, it will request an exclusive lock with T (try 1CB
) set. It will continue to deallocate the inode if the lock is granted. If the lock is not granted it will result in the node(s) which were preventing the grant of the lock marking their glock(s) with the D (demote) flag, which is checked at →drop_inode
() time in order to ensure that the deallocation is not forgotten.
This means that inodes that have zero link count but are still open will be deallocated by the node on which the final close
() occurs. Also, at the same time as the inode’s link count is decremented to zero the inode is marked as being in the special state of having zero link count but still in use in the resource group bitmap. This functions like the ext3 file system3’s orphan list in that it allows any subsequent reader of the bitmap to know that there is potentially space that might be reclaimed, and to attempt to reclaim it.
9.6. Glock tracepoints
The tracepoints are also designed to be able to confirm the correctness of the cache control by combining them with the blktrace
output and with knowledge of the on-disk layout. It is then possible to check that any given I/O has been issued and completed under the correct lock, and that no races are present.
The gfs2_glock_state_change
tracepoint is the most important one to understand. It tracks every state change of the glock from initial creation right through to the final demotion which ends with gfs2_glock_put
and the final NL to unlocked transition. The l (locked) glock flag is always set before a state change occurs and will not be cleared until after it has finished. There are never any granted holders (the H glock holder flag) during a state change. If there are any queued holders, they will always be in the W (waiting) state. When the state change is complete then the holders may be granted which is the final operation before the l glock flag is cleared.
The gfs2_demote_rq
tracepoint keeps track of demote requests, both local and remote. Assuming that there is enough memory on the node, the local demote requests will rarely be seen, and most often they will be created by umount
or by occasional memory reclaim. The number of remote demote requests is a measure of the contention between nodes for a particular inode or resource group.
The gfs2_glock_lock_time
tracepoint provides information about the time taken by requests to the DLM. The blocking (b
) flag was introduced into the glock specifically to be used in combination with this tracepoint.
When a holder is granted a lock, gfs2_promote
is called, this occurs as the final stages of a state change or when a lock is requested which can be granted immediately due to the glock state already caching a lock of a suitable mode. If the holder is the first one to be granted for this glock, then the f (first) flag is set on that holder. This is currently used only by resource groups.
9.7. Bmap tracepoints
Block mapping is a task central to any file system. GFS2 uses a traditional bitmap-based system with two bits per block. The main purpose of the tracepoints in this subsystem is to allow monitoring of the time taken to allocate and map blocks.
The gfs2_bmap
tracepoint is called twice for each bmap operation: once at the start to display the bmap request, and once at the end to display the result. This makes it easy to match the requests and results together and measure the time taken to map blocks in different parts of the file system, different file offsets, or even of different files. It is also possible to see what the average extent sizes being returned are in comparison to those being requested.
The gfs2_rs
tracepoint traces block reservations as they are created, used, and destroyed in the block allocator.
To keep track of allocated blocks, gfs2_block_alloc
is called not only on allocations, but also on freeing of blocks. Since the allocations are all referenced according to the inode for which the block is intended, this can be used to track which physical blocks belong to which files in a live file system. This is particularly useful when combined with blktrace
, which will show problematic I/O patterns that may then be referred back to the relevant inodes using the mapping gained by means this tracepoint.
Direct I/O (iomap
) is an alternative cache policy which allows file data transfers to happen directly between disk and the user’s buffer. This has benefits in situations where cache hit rate is expected to be low. Both gfs2_iomap_start
and gfs2_iomap_end
tracepoints trace these operations and can be used to keep track of mapping using Direct I/O, the positions on the file system of the Direct I/O along with the operation type.
9.8. Log tracepoints
The tracepoints in this subsystem track blocks being added to and removed from the journal (gfs2_pin
), as well as the time taken to commit the transactions to the log (gfs2_log_flush
). This can be very useful when trying to debug journaling performance issues.
The gfs2_log_blocks
tracepoint keeps track of the reserved blocks in the log, which can help show if the log is too small for the workload, for example.
The gfs2_ail_flush
tracepoint is similar to the gfs2_log_flush
tracepoint in that it keeps track of the start and end of flushes of the AIL list. The AIL list contains buffers which have been through the log, but have not yet been written back in place and this is periodically flushed in order to release more log space for use by the file system, or when a process requests a sync
or fsync
.
9.9. Glock statistics
GFS2 maintains statistics that can help track what is going on within the file system. This allows you to spot performance issues.
GFS2 maintains two counters:
-
dcount
, which counts the number of DLM operations requested. This shows how much data has gone into the mean/variance calculations. -
qcount
, which counts the number ofsyscall
level operations requested. Generallyqcount
will be equal to or greater thandcount
.
In addition, GFS2 maintains three mean/variance pairs. The mean/variance pairs are smoothed exponential estimates and the algorithm used is the one used to calculate round trip times in network code.
The mean and variance pairs maintained in GFS2 are not scaled, but are in units of integer nanoseconds.
- srtt/srttvar: Smoothed round trip time for non-blocking operations
- srttb/srttvarb: Smoothed round trip time for blocking operations
- irtt/irttvar: Inter-request time (for example, time between DLM requests)
A non-blocking request is one which will complete right away, whatever the state of the DLM lock in question. That currently means any requests when (a) the current state of the lock is exclusive (b) the requested state is either null or unlocked or (c) the "try lock" flag is set. A blocking request covers all the other lock requests.
Larger times are better for IRTTs, whereas smaller times are better for the RTTs.
Statistics are kept in two sysfs
files:
-
The
glstats
file. This file is similar to theglocks
file, except that it contains statistics, with one glock per line. The data is initialized from "per cpu" data for that glock type for which the glock is created (aside from counters, which are zeroed). This file may be very large. -
The
lkstats
file. This contains "per cpu" stats for each glock type. It contains one statistic per line, in which each column is a cpu core. There are eight lines per glock type, with types following on from each other.
9.10. References
For more information about tracepoints and the GFS2 glocks
file, see the following resources:
- For information about glock internal locking rules, see https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/filesystems/gfs2-glocks.rst.
- For information about event tracing, see https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/trace/events.rst.
-
For information about the
trace-cmd
utility, see http://lwn.net/Articles/341902/.
Chapter 10. Monitoring and analyzing GFS2 file systems using Performance Co-Pilot (PCP)
Performance Co-Pilot (PCP) can help with monitoring and analyzing GFS2 file systems. Monitoring of GFS2 file systems in PCP is provided by the GFS2 PMDA module in Red Hat Enterprise Linux which is available through the pcp-pmda-gfs2
package.
The GFS2 PMDA provides a number of metrics given by the GFS2 statistics provided in the debugfs
subsystem. When installed, the PMDA exposes values given in the glocks
, glstats
, and sbstats
files. These report sets of statistics on each mounted GFS2 filesystem. The PMDA also makes use of the GFS2 kernel tracepoints exposed by the Kernel Function Tracer (ftrace
).
10.1. Installing the GFS2 PMDA
In order to operate correctly, The GFS2 PMDA requires that the debugfs
file system is mounted. If the debugfs
file system is not mounted, run the following commands before installing the GFS2 PMDA:
# mkdir /sys/kernel/debug # mount -t debugfs none /sys/kernel/debug
The GFS2 PMDA is not enabled as part of the default installation. In order to make use of GFS2 metric monitoring through PCP you must enable it after installation.
Run the following commands to install PCP and enable the GFS2 PMDA. Note that the PMDA install script must be run as root.
# dnf install pcp pcp-pmda-gfs2 # cd /var/lib/pcp/pmdas/gfs2 # ./Install Updating the Performance Metrics Name Space (PMNS) ... Terminate PMDA if already installed ... Updating the PMCD control file, and notifying PMCD ... Check gfs2 metrics have appeared ... 346 metrics and 255 values
10.2. Displaying information about the available performance metrics with the pminfo tool
The pminfo
tool displays information about the available performance metrics. The following examples show different GFS2 metrics you can display with this tool.
10.2.1. Examining the number of glock structures that currently exist per file system
The GFS2 glock metrics give insights to the number of glock structures currently incore for each mounted GFS2 file system and their locking states. In GFS2, a glock is a data structure that brings together the DLM and caching into a single state machine. Each glock has a 1:1 mapping with a single DLM lock and provides caching for the lock states so that repetitive operations carried out on a single node do not have to repeatedly call the DLM, reducing unnecessary network traffic.
The following pminfo
command displays a list of the number of glocks per mounted GFS2 file system by their lock mode.
# pminfo -f gfs2.glocks
gfs2.glocks.total
inst [0 or "afc_cluster:data"] value 43680
inst [1 or "afc_cluster:bin"] value 2091
gfs2.glocks.shared
inst [0 or "afc_cluster:data"] value 25
inst [1 or "afc_cluster:bin"] value 25
gfs2.glocks.unlocked
inst [0 or "afc_cluster:data"] value 43652
inst [1 or "afc_cluster:bin"] value 2063
gfs2.glocks.deferred
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.glocks.exclusive
inst [0 or "afc_cluster:data"] value 3
inst [1 or "afc_cluster:bin"] value 3
10.2.2. Examining the number of glock structures that exist per file system by type
The GFS2 glstats metrics give counts of each type of glock which exist for each files ystem, a large number of these will normally be of either the inode (inode and metadata) or resource group (resource group metadata) type.
The following pminfo
command displays a list of the number of each type of Glock per mounted GFS2 file system.
# pminfo -f gfs2.glstats
gfs2.glstats.total
inst [0 or "afc_cluster:data"] value 43680
inst [1 or "afc_cluster:bin"] value 2091
gfs2.glstats.trans
inst [0 or "afc_cluster:data"] value 3
inst [1 or "afc_cluster:bin"] value 3
gfs2.glstats.inode
inst [0 or "afc_cluster:data"] value 17
inst [1 or "afc_cluster:bin"] value 17
gfs2.glstats.rgrp
inst [0 or "afc_cluster:data"] value 43642
inst [1 or "afc_cluster:bin"] value 2053
gfs2.glstats.meta
inst [0 or "afc_cluster:data"] value 1
inst [1 or "afc_cluster:bin"] value 1
gfs2.glstats.iopen
inst [0 or "afc_cluster:data"] value 16
inst [1 or "afc_cluster:bin"] value 16
gfs2.glstats.flock
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.glstats.quota
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.glstats.journal
inst [0 or "afc_cluster:data"] value 1
inst [1 or "afc_cluster:bin"] value 1
10.2.3. Checking the number of glock structures that are in a wait state
The most important holder flags are H (holder: indicates that requested lock is granted) and W (wait: set while waiting for request to complete). These flags are set on granted lock requests and queued lock requests, respectively.
The following pminfo
command displays a list of the number of glocks with the Wait (W) holder flag for each mounted GFS2 file system.
# pminfo -f gfs2.holders.flags.wait
gfs2.holders.flags.wait
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
If you do see a number of waiting requests queued on a resource group lock there may be a number of reasons for this. One is that there are a large number of nodes compared to the number of resource groups in the file system. Another is that the file system may be very nearly full (requiring, on average, longer searches for free blocks). The situation in both cases can be improved by adding more storage and using the gfs2_grow
command to expand the file system.
10.2.4. Checking file system operation latency using the kernel tracepoint based metrics
The GFS2 PMDA supports collecting of metrics from the GFS2 kernel tracepoints. By default the reading of these metrics is disabled. Activating these metrics turns on the GFS2 kernel tracepoints when the metrics are collected in order to populate the metric values. This could have a small effect on performance throughput when these Kernel Tracepoint metrics are enabled.
PCP provides the pmstore
tool, which allows you to modify PMDA settings based on metric values. The gfs2.control.*
metrics allow the toggling of GFS2 kernel tracepoints. The following example uses the pmstore
command to enable all of the GFS2 kernel tracepoints.
# pmstore gfs2.control.tracepoints.all 1
gfs2.control.tracepoints.all old value=0 new value=1
When this command is run, the PMDA switches on all of the GFS2 tracepoints in the debugfs
file system. The "Complete Metric List" table in Complete listing of available metrics for GFS2 in PCP explains each of the control tracepoints and their usage, An explanation on the effect of each control tracepoint and its available options is also available through the help switch in pminfo
.
The GFS2 promote metrics count the number of promote requests on the file system. These requests are separated by the number of requests that have occurred on the first attempt and “others” which are granted after their initial promote request. A drop in the number of first time promotes with a rise in “other” promotes can indicate issues with file contention.
The GFS2 demote request metrics, like the promote request metrics, count the number of demote requests which occur on the file system. These, however, are also split between requests that have come from the current node and requests that have come from other nodes on the system. A large number of demote requests from remote nodes can indicate contention between two nodes for a given resource group.
The pminfo
tool displays information about the available performance metrics. This procedure displays a list of the number of glocks with the Wait (W) holder flag for each mounted GFS2 file system. The following pminfo
command displays a list of the number of glocks with the Wait (W) holder flag for each mounted GFS2 file system.
# pminfo -f gfs2.latency.grant.all gfs2.latency.demote.all
gfs2.latency.grant.all
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.latency.demote.all
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
It is a good idea to determine the general values observed when the workload is running without issues to be able to notice changes in performance when these values differ from their normal range.
For example, you might notice a change in the number of promote requests waiting to complete rather than completing on first attempt, which the output from following command would allow you to determine.
# pminfo -f gfs2.latency.grant.all gfs2.latency.demote.all
gfs2.tracepoints.promote.other.null_lock
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.tracepoints.promote.other.concurrent_read
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.tracepoints.promote.other.concurrent_write
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.tracepoints.promote.other.protected_read
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.tracepoints.promote.other.protected_write
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.tracepoints.promote.other.exclusive
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
The output from following command would allow you to determine a large increase in remote demote requests (especially if from other cluster nodes).
# pminfo -f gfs2.tracepoints.demote_rq.requested
gfs2.tracepoints.demote_rq.requested.remote
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
gfs2.tracepoints.demote_rq.requested.local
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
The output from the following command could indicate an unexplained increase in log flushes.
# pminfo -f gfs2.tracepoints.log_flush.total]
gfs2.tracepoints.log_flush.total
inst [0 or "afc_cluster:data"] value 0
inst [1 or "afc_cluster:bin"] value 0
10.3. Complete listing of available metrics for GFS2 in PCP
The following table describes the full list of performance metrics given by the pcp-pmda-gfs2
package for GFS2 file systems.
Metric Name | Description |
---|---|
|
Metrics regarding the information collected from the glock stats file ( |
| Range of metrics counting the number of glocks that exist with the given glocks flags |
|
Metrics regarding the information collected from the glock stats file ( |
| Range of metrics counting the number of glocks holders with the given holder flags |
|
Timing metrics regarding the information collected from the superblock stats file ( |
|
Metrics regarding the information collected from the glock stats file ( |
|
A derived metric making use of the data from both the |
|
A derived metric making use of the data from both the |
|
A derived metric making use of the data from the |
|
A derived metric making use of the data from the |
|
Metrics regarding the output from the GFS2 |
|
Configuration metrics which are used to switch on or off metric recording in the PMDA. Conrol metricsare toggled by means of the |
10.4. Performing minimal PCP setup to gather file system data
This procedure outlines instructions on how to install a minimal PCP setup to collect statistics on Red Hat Enterprise Linux. This setup involves adding the minimum number of packages on a production system needed to gather data for further analysis.
The resulting tar.gz
archive of the pmlogger
output can be analyzed by using further PCP tools and can be compared with other sources of performance information.
Procedure
Install the required PCP packages.
# dnf install pcp pcp-pmda-gfs2
Activate the GFS2 module for PCP.
# cd /var/lib/pcp/pmdas/gfs2 # ./Install
Start both the
pmcd
andpmlogger
services.# systemctl start pmcd.service # systemctl start pmlogger.service
- Perform operations on the GFS2 file system.
Stop both the
pmcd
andpmlogger
services.# systemctl stop pmcd.service # systemctl stop pmlogger.service
Collect the output and save it to a
tar.gz
file named based on the host name and the current date and time.# cd /var/log/pcp/pmlogger # tar -czf $(hostname).$(date+%F-%Hh%M).pcp.tar.gz $(hostname)