Installing the device-mapper-multipath RPM package could have caused the default installed /etc/kpartx.conf configuration file to be modified. As a consequence, verifying the package with the rpm -V command failed due to the /etc/kpartx.conf file's size, md5sum, and last modification time having been changed. This update adds flags to the device-mapper-multipath package's spec file that inform rpm that the size, md5sum and modification time of the /etc/kpartx.conf file may change, with the result that verifying the package now succeeds in this situation.

When the system boots, multipath reads the WWID (World Wide Identifier) of the storage device on each path, and creates a binding from each WWID to an mpath[n] device name. By default, these bindings are written to the /var/lib/multipath/bindings file, which is referenced in all future scans to ensure that the binding from an mpath[n] to a storage device WWID remains constant for the life of the system. When the /var directory was mounted on a separate file system from root ("/"), then /var could have been mounted after multipath had created the initial bindings file. In this situation, the next time multipath ran, it did not find a bindings file and therefore proceeded to create new bindings, which it wrote to the /var/lib/multipath/bindings file.

However, if the scan order differed, then the bindings would change, thus causing one or more multipath[n] names to point to different storage devices. This could have resulted in: inconsistent multipath maps; multiple maps with the same WWID; maps with the wrong paths; and potentially data loss. Note that the possibility that the scan order can potentially differ is greater in an environment where LUNs (Logical Unit Numbers) are added or removed after the system boots.

The multipath.conf configuration file now includes a default bindings_file directive. On installations on which the /var/lib/multipath/bindings file is not located on the same device as the root file system, then this directive should specify a custom path to a bindings file on the same device as the root file system, for multipath to use instead of the default. By setting this directive, multipath will use the same bindings during boot as it does during normal operation, multipath devices will remain consistent, and the possibility of data loss is avoided.

For more information on the bindings_file directive, refer to the /usr/share/doc/device-mapper-multipath-0.4.5/multipath.conf.annotated file.

Multipath is unable to reconfigure a multipath device while the multipathd service is running, or add new paths to the device, because the kernel does not allow multipath to reload a table with a failed path in it. In order to remove a failed path from the device table, the following procedure should be followed as a workaround:

Running the multipath command while the multipathd service is stopped causes multipath to remove the failed path from the device table when it reconfigures, thus allowing the new path to be added. However, when the failed path comes back online, it will no longer be part of the multipath device. The multipath command must be manually run again in order to reintegrate this device. Note that if there are no failed paths, it is not necessary to stop the multipathd service before running multipath.

When using the sky2 driver for Marvell Ethernet adapters with the following device:

...packet transmission may time out with the following message being written to netconsole: sky2 eth1: tx timeout.

This is a known issue: after receiving these messages, the system must be rebooted in order to fix the packet transmission issues.

When dm-multipath is used on a storage device that implements ALUA, and group-by-prio is enabled, then the path groups are established when the device is configured. The paths with the same priority are grouped together, the group priority is calculated as the sum of the path priorities, and the path group with the highest sum is selected for I/O. If a path's priority changes, the group priority is re-calculated, and the active path group may change. The path grouping is not changed, even though some members of the group may now have different priorities. If you wish to re-establish the path grouping after a change, then you must enter the command

multipathd -k"reconfigure"

When dm-multipath is used on a storage device that implements ALUA, and group-by-prio is enabled, then the path groups are established when the device is configured. The paths with the same priority are grouped together, the group priority is calculated as the sum of the path priorities, and the path group with the highest sum is selected for I/O. If a path's priority changes, the group priority is re-calculated, and the active path group may change. The path grouping is not changed, even though some members of the group may now have different priorities. If you wish to re-establish the path grouping after a change, then you must enter the command

multipathd -k"reconfigure"

Broken dependencies in the up2date-4.8.1-33.el4_8.9 package may result in an update of the up2date package failing. Consequently, if the up2date-4.8.1-33.el4_8.9, package is installed on a system the new up2date package must be manually downloaded from the Red Hat Network and installed using the rpm command.

Japanese language JP106 keyboards will not function correctly when booting into Rescue Mode on Red Hat Enterprise Linux 4.9.

Updating all packages from Red Hat Enterprise Linux 4.8 to Red Hat Enterprise Linux 4.9 on multilib architectures may fail with dependency issues for the openmpi-libs package. To work around this issue, use the following commands to update the compat-dapl package before updating the remaining packages:

up2date compat-dapl
up2date -fu

Some x86 64-bit systems may hang during boot when the noapic debug kernel parameter is used.

As a result of N_Port ID Virtualization (NPIV) support added in Red Hat Enterprise Linux 4.8 on s390x architectures, the permanent_port_name sysfs attribute is no longer included. This attribute was used (primarily for debugging purposes) to differentiate the use of NPIV Logical Unit Numbers (LUNs) from within Linux. In the absence of this attribute, system administrators should refer to the Hardware Management Console / Support Element (HMC/SE) to find the virtual port address on an NPIV-enabled system.

The Logical Volume Manager in Red Hat Enterprise Linux 4.9 reports file descriptor leaks, resulting in the following error returned to the installation output:

				File descriptor NUM (socket:XXXX) leaked on lvm invocation.

This message can be safely ignored.

When installing Red Hat Enterprise Linux 4 through an Network File System (NFS) server, the installer is unable to correctly close the NFS mount points. This might cause the NFS server to misbehave. In these cases Red Hat suggests the use of an HTTP server for installations.

On systems where the BIOS is able to do both legacy (acpiphp) and native (pciehp) PCI hotplugging, it is necessary for the administrator to choose a preferred method and explicitly prevent Red Hat Enterprise Linux 4 from loading the module for the undesired method. This is done by blacklisting the undesired module in /etc/modprobe.conf.

Red Hat discourages the use of quota on EXT3 file systems. This is because in some cases, doing so can cause a deadlock.

Testing has revealed that kjournald can sometimes block some EXT3-specific callouts that are used when quota is running. As such, Red Hat does not plan to fix this issue in Red Hat Enterprise Linux 4, as the modifications required would be too invasive.

Note that this issue is not present in Red Hat Enterprise Linux 5.

The Desktop Sharing connection icon displays its context menu when you double-click it, not when you right-click it. All other icons display their context menus when you right-click on them.

If the ib_ehca InfiniBand driver is loaded in port auto-detection mode (using module parameter nr_ports=-1), the IP-over-InfiniBand network interfaces (ibX) might become available too late. When this occurs, the ifup ibX command issued from the openibd startup script will fail; consequently, the ibX interface will not become available.

When this occurs, use the command rcnetwork restart to fix the problem.

In the IBM Redbook "Implementing InfiniBand in IBM System p (SG247351) manual, Table 6-3 (on page 220 of the PDF version) describes debug code bit definitions, where several HCA error indicator bits are also described.

Note that with eHCA2 adapters, bits 46 and 47 of these error indicator bits might return false positives.

On HP ICH10 workstations, audio is only enabled through the front 3.5mm jacks. As such, to receive any audio output or use recording, you should plug in your headphones, speakers, or microphones to the front jacks. At present, the rear jacks, internal speaker, and master volume for this workstation do not work.

With this update, the default PCI detection and ordering mode for the following models have changed:

These models use a device scanning and enumeration mode which is not the default for Red Hat Enterprise Linux 4 or 5. The mode used by these HP Proliant models could result in add-on cards being detected and added prior to onboard/internal devices. This unexpected ordering could cause difficulties when installing new instances of Red Hat Enterprise Linux, adding hardware, and maintenance.

The numbering of network interface cards (NIC) for the aforementioned HP Proliant models may change when they are updated with the Red Hat Enterprise Linux 4.7 kernel. The installer changes NIC numbering if the HWADDR=MAC ADDRESS parameter is not defined in /etc/sysconfig/network-scripts/ifcfg-eth[X] for each installed NICs. As such, Red Hat recommends that you ensure this parameter is defined in order to avoid any problems arising from an unexpected NIC enumeration.

In addition, to avoid any NIC enumeration changes after updating these HP Proliant models to Red Hat Enterprise Linux 4.7, add the kernel boot parameter pci=nobfsort to /boot/grub/grub.conf.

When a volume group contains a mirror or snapshot, issuing the lvchange command with a volume group parameter may result in the following error messages:

Unable to change mirror log LV fail_secondary_mlog directly
Unable to change mirror image LV fail_secondary_mimage_0 directly
Unable to change mirror image LV fail_secondary_mimage_1 directly

These messages can be safely ignored.

Dell PowerEdge SC1435s systems may hang during boot-up. To avoid this, edit the terminal line in grub.conf and replace the string serial console with console serial.

The updated ixgbe driver does not support the Intel 82598AT (Copper Pond 10GbE).

Red Hat Enterprise Linux 4.9 can detect online growing or shrinking of an underlying block device. However, there is no method to automatically detect that a device has changed size, so manual steps are required to recognize this and resize any file systems which reside on the given device(s). When a resized block device is detected, a message like the following will appear in the system logs:

VFS: busy inodes on changed media or resized disk sdi

If the block device was grown, then this message can be safely ignored. However, if the block device was shrunk without shrinking any data set on the block device first, the data residing on the device may be corrupted.

It is only possible to do an online resize of a filesystem that was created on the entire LUN (or block device). If there is a partition table on the block device, then the file system will have to be unmounted to update the partition table.

There is a known memory leak with the res_n* family of resolver routines (i.e. res_nquery, res_nsearch and res_nmkquery). Programs that use these functions will leak memory over time. It has been fixed in newer versions of glibc, however, the fix is too invasive to be applied to Red Hat Enterprise Linux 4. Programs that use these functions may need to be restarted occasionally to free memory.

The number of devices that can be handled during installation of Red Hat Enterprise Linux 4 depends on the size of the installation initrd image. Therefore, in situations where there are many devices attached to a machine (such as heavily populated Fibre Channel setups) installation will not be possible unless number of visible devices is reduced.

During installation anaconda may not remove all the Logical Volume Manager (LVM) metadata that exists on a system prior to installation. This extra metadata may cause LVM tools to report missing volume groups or logical volumes after installation. To work around this issue, remove the stale LVM metadata after the installation is complete.

multipath does not silence the error messages printed by any of it's callout programs. Therefore, if multipath is run when paths are down, various error messages may be displayed. The messages that are displayed depend on the specific callout programs that multipath is using. For example, if multipath is run while there are failed scsi devices, scsi_id will print

lt;H>:<B>:<T>:<L>:Unable to get INQUIRY vpd 1 page 0x0.
lt;H>:<B>:<T>:<L>:sg_io failed status 0x0 0x1 0x0 0x0

Or, if multipath -ll is run while an EMC CLARiiON is down, the mpath_prio_emc priority callout will print query command indicates error

On some SGI Altix systems that feature the IOC4 multi-function device, you may encounter problems when using attached IDE devices (such as CD-ROM drives). This is caused by a bug in the sgiioc4 IDE driver, which prevents some devices from being detected properly on system boot.

You can work around this bug by manually loading the driver, which in turn allows attached IDE devices to be detected properly. To do so, run the following command as root:

/sbin/modprobe sgiioc4