Chapter 17. Configuring a Linux instance on 64-bit IBM Z

Direct Access Storage Devices (DASDs) are a type of storage commonly used with 64-bit IBM Z. The following example is how to set a DASD online, format it, and make the change persistent.

Verify that the device is attached or linked to the Linux system if running under z/VM.

CP ATTACH EB1C TO *

To link a mini disk to which you have access, run the following commands:

CP LINK RHEL7X 4B2E 4B2E MR
DASD 4B2E LINKED R/W

This section contains information about setting a DASD online.

For instructions on how to set a DASD online persistently, see Persistently setting DASDs online.

Once the disk is online, change back to the /root directory and low-level format the device. This is only required once for a DASD during its entire lifetime:

# cd /root
# dasdfmt -b 4096 -d cdl -p /dev/disk/by-path/ccw-0.0.4b2e
Drive Geometry: 10017 Cylinders * 15 Heads =  150255 Tracks

I am going to format the device /dev/disk/by-path/ccw-0.0.4b2e in the following way:
Device number of device : 0x4b2e
Labelling device        : yes
Disk label              : VOL1
Disk identifier         : 0X4B2E
Extent start (trk no)   : 0
Extent end (trk no)     : 150254
Compatible Disk Layout  : yes
Blocksize               : 4096

--->> ATTENTION! <<---
All data of that device will be lost.
Type "yes" to continue, no will leave the disk untouched: yes
cyl    97 of  3338 |#----------------------------------------------|   2%

When the progress bar reaches the end and the format is complete, dasdfmt prints the following output:

Rereading the partition table...
Exiting...

Now, use fdasd to partition the DASD. You can create up to three partitions on a DASD. In our example here, we create one partition spanning the whole disk:

# fdasd -a /dev/disk/by-path/ccw-0.0.4b2e
reading volume label ..: VOL1
reading vtoc ..........: ok

auto-creating one partition for the whole disk...
writing volume label...
writing VTOC...
rereading partition table...

After a (low-level formatted) DASD is online, it can be used like any other disk under Linux. For example, you can create file systems, LVM physical volumes, or swap space on its partitions, for example /dev/disk/by-path/ccw-0.0.4b2e-part1. Never use the full DASD device (dev/dasdb) for anything but the commands dasdfmt and fdasd. If you want to use the entire DASD, create one partition spanning the entire drive as in the fdasd example above.

To add additional disks later without breaking existing disk entries in, for example, /etc/fstab, use the persistent device symbolic links under /dev/disk/by-path/.

The above instructions described how to activate DASDs dynamically in a running system. However, such changes are not persistent and do not survive a reboot. Making changes to the DASD configuration persistent in your Linux system depends on whether the DASDs belong to the root file system. Those DASDs required for the root file system need to be activated very early during the boot process by the initramfs to be able to mount the root file system.

The cio_ignore commands are handled transparently for persistent device configurations and you do not need to free devices from the ignore list manually.

The file you have to modify to add DASDs that are part of the root file system has changed in Red Hat Enterprise Linux 8. Instead of editing the /etc/zipl.conf file, the new file to be edited, and its location, may be found by running the following commands:

# machine_id=$(cat /etc/machine-id)
# kernel_version=$(uname -r)
# ls /boot/loader/entries/$machine_id-$kernel_version.conf

There is one boot option to activate DASDs early in the boot process: rd.dasd=. This option takes a Direct Access Storage Device (DASD) adapter device bus identifier. For multiple DASDs, specify the parameter multiple times, or use a comma separated list of bus IDs. To specify a range of DASDs, specify the first and the last bus ID. Below is an example of the /boot/loader/entries/4ab74e52867b4f998e73e06cf23fd761-4.18.0-80.el8.s390x.conf file for a system that uses physical volumes on partitions of two DASDs for an LVM volume group vg_devel1 that contains a logical volume lv_root for the root file system.

title Red Hat Enterprise Linux (4.18.0-80.el8.s390x) 8.0 (Ootpa)
version 4.18.0-80.el8.s390x
linux /boot/vmlinuz-4.18.0-80.el8.s390x
initrd /boot/initramfs-4.18.0-80.el8.s390x.img
options root=/dev/mapper/vg_devel1-lv_root crashkernel=auto rd.dasd=0.0.0200 rd.dasd=0.0.0207 rd.lvm.lv=vg_devel1/lv_root rd.lvm.lv=vg_devel1/lv_swap cio_ignore=all,!condev rd.znet=qeth,0.0.0a00,0.0.0a01,0.0.0a02,layer2=1,portno=0
id rhel-20181027190514-4.18.0-80.el8.s390x
grub_users $grub_users
grub_arg --unrestricted
grub_class kernel

To add another physical volume on a partition of a third DASD with device bus ID 0.0.202b. To do this, add rd.dasd=0.0.202b to the parameters line of your boot kernel in /boot/loader/entries/4ab74e52867b4f998e73e06cf23fd761-4.18.0-32.el8.s390x.conf:

title Red Hat Enterprise Linux (4.18.0-80.el8.s390x) 8.0 (Ootpa)
version 4.18.0-80.el8.s390x
linux /boot/vmlinuz-4.18.0-80.el8.s390x
initrd /boot/initramfs-4.18.0-80.el8.s390x.img
options root=/dev/mapper/vg_devel1-lv_root crashkernel=auto rd.dasd=0.0.0200 rd.dasd=0.0.0207 rd.dasd=0.0.202b rd.lvm.lv=vg_devel1/lv_root rd.lvm.lv=vg_devel1/lv_swap cio_ignore=all,!condev rd.znet=qeth,0.0.0a00,0.0.0a01,0.0.0a02,layer2=1,portno=0
id rhel-20181027190514-4.18.0-80.el8.s390x
grub_users $grub_users
grub_arg --unrestricted
grub_class kernel

Run zipl to apply the changes of the configuration file for the next IPL:

# zipl -V
Using config file '/etc/zipl.conf'
Using BLS config file '/boot/loader/entries/4ab74e52867b4f998e73e06cf23fd761-4.18.0-80.el8.s390x.conf'
Target device information
  Device..........................: 5e:00
  Partition.......................: 5e:01
  Device name.....................: dasda
  Device driver name..............: dasd
  DASD device number..............: 0201
  Type............................: disk partition
  Disk layout.....................: ECKD/compatible disk layout
  Geometry - heads................: 15
  Geometry - sectors..............: 12
  Geometry - cylinders............: 13356
  Geometry - start................: 24
  File system block size..........: 4096
  Physical block size.............: 4096
  Device size in physical blocks..: 262152
Building bootmap in '/boot'
Building menu 'zipl-automatic-menu'
Adding #1: IPL section '4.18.0-80.el8.s390x' (default)
  initial ramdisk...: /boot/initramfs-4.18.0-80.el8.s390x.img
  kernel image......: /boot/vmlinuz-4.18.0-80.el8.s390x
  kernel parmline...: 'root=/dev/mapper/vg_devel1-lv_root crashkernel=auto rd.dasd=0.0.0200 rd.dasd=0.0.0207 rd.dasd=0.0.202b rd.lvm.lv=vg_devel1/lv_root rd.lvm.lv=vg_devel1/lv_swap cio_ignore=all,!condev rd.znet=qeth,0.0.0a00,0.0.0a01,0.0.0a02,layer2=1,portno=0'
  component address:
    kernel image....: 0x00010000-0x0049afff
    parmline........: 0x0049b000-0x0049bfff
    initial ramdisk.: 0x004a0000-0x01a26fff
    internal loader.: 0x0000a000-0x0000cfff
Preparing boot menu
  Interactive prompt......: enabled
  Menu timeout............: 5 seconds
  Default configuration...: '4.18.0-80.el8.s390x'
Preparing boot device: dasda (0201).
Syncing disks...
Done.

Direct Access Storage Devices (DASDs) that are not part of the root file system, that is, data disks, are persistently configured in the /etc/dasd.conf file. This file contains one DASD per line, where each line begins with the DASD’s bus ID.

When adding a DASD to the /etc/dasd.conf file, use key-value pairs to specify the options for each entry. Separate the key and its value with an equal (=) sign. When adding multiple options, use a space or a tab to separate each option.

0.0.0207
0.0.0200 use_diag=1 readonly=1

Changes to the /etc/dasd.conf file take effect after a system reboot or after a new DASD is dynamically added by changing the system’s I/O configuration (that is, the DASD is attached under z/VM).

Alternatively, to activate a DASD that you have added to the /etc/dasd.conf file, complete the following steps:

The only file you have to modify for adding FCP LUNs that are part of the root file system has changed in Red Hat Enterprise Linux 9. Instead of editing the /etc/zipl.conf file, the new file to be edited, and its location, may be found by running the following commands:

# machine_id=$(cat /etc/machine-id)
# kernel_version=$(uname -r)
# ls /boot/loader/entries/$machine_id-$kernel_version.conf

Red Hat Enterprise Linux provides a parameter to activate FCP LUNs early in the boot process: rd.zfcp=. The value is a comma-separated list containing the FCP device bus ID, the target WWPN as 16 digit hexadecimal number prefixed with 0x, and the FCP LUN prefixed with 0x and padded with zeroes to the right to have 16 hexadecimal digits.

The WWPN and FCP LUN values are only necessary if the zFCP device is not configured in NPIV mode, when auto LUN scanning is disabled by the zfcp.allow_lun_scan=0 kernel module parameter or when installing RHEL-9.0 or older releases. Otherwise they can be omitted, for example, rd.zfcp=0.0.4000. Below is an example of the /boot/loader/entries/4ab74e52867b4f998e73e06cf23fd761-5.14.0-55.el9.s390x.conf file for a system that uses a physical volume on a partition of an FCP-attached SCSI disk, with two paths, for an LVM volume group vg_devel1 that contains a logical volume lv_root for the root file system.

title Red Hat Enterprise Linux (5.14.0-55.el9.s390x) 9.0 (Plow)
version 5.14.0-55.el9.s390x
linux /boot/vmlinuz-5.14.0-55.el9.s390x
initrd /boot/initramfs-5.14.0-55.el9.s390x.img
options root=/dev/mapper/vg_devel1-lv_root crashkernel=auto rd.zfcp=0.0.fc00,0x5105074308c212e9,0x401040a000000000 rd.zfcp=0.0.fcd0,0x5105074308c2aee9,0x401040a000000000 rd.lvm.lv=vg_devel1/lv_root rd.lvm.lv=vg_devel1/lv_swap cio_ignore=all,!condev rd.znet=qeth,0.0.0a00,0.0.0a01,0.0.0a02,layer2=1,portno=0
id rhel-20181027190514-5.14.0-55.el9.s390x
grub_users $grub_users
grub_arg --unrestricted
grub_class kernel

title Red Hat Enterprise Linux (5.14.0-55.el9.s390x) 9.0 (Plow)
version 5.14.0-55.el9.s390x
linux /boot/vmlinuz-5.14.0-55.el9.s390x
initrd /boot/initramfs-5.14.0-55.el9.s390x.img
options root=/dev/mapper/vg_devel1-lv_root crashkernel=auto rd.zfcp=0.0.fc00,0x5105074308c212e9,0x401040a000000000 rd.zfcp=0.0.fcd0,0x5105074308c2aee9,0x401040a000000000 rd.zfcp=0.0.fc00,0x5105074308c212e9,0x401040a300000000 rd.zfcp=0.0.fcd0,0x5105074308c2aee9,0x401040a300000000 rd.lvm.lv=vg_devel1/lv_root rd.lvm.lv=vg_devel1/lv_swap cio_ignore=all,!condev rd.znet=qeth,0.0.0a00,0.0.0a01,0.0.0a02,layer2=1,portno=0
id rhel-20181027190514-5.14.0-55.el9.s390x
grub_users $grub_users
grub_arg --unrestricted
grub_class kernel

# zipl -V
Using config file '/etc/zipl.conf'
Using BLS config file '/boot/loader/entries/4ab74e52867b4f998e73e06cf23fd761-5.14.0-55.el9.s390x.conf'
Run /lib/s390-tools/zipl_helper.device-mapper /boot
Target device information
Device..........................: fd:00
Partition.......................: fd:01
Device name.....................: dm-0
Device driver name..............: device-mapper
Type............................: disk partition
Disk layout.....................: SCSI disk layout
Geometry - start................: 2048
File system block size..........: 4096
Physical block size.............: 512
Device size in physical blocks..: 10074112
Building bootmap in '/boot/'
Building menu 'zipl-automatic-menu'
Adding #1: IPL section '5.14.0-55.el9.s390x' (default)
kernel image......: /boot/vmlinuz-5.14.0-55.el9.s390x
kernel parmline...: 'root=/dev/mapper/vg_devel1-lv_root crashkernel=auto rd.zfcp=0.0.fc00,0x5105074308c212e9,0x401040a000000000 rd.zfcp=0.0.fcd0,0x5105074308c2aee9,0x401040a000000000 rd.zfcp=0.0.fc00,0x5105074308c212e9,0x401040a300000000 rd.zfcp=0.0.fcd0,0x5105074308c2aee9,0x401040a300000000 rd.lvm.lv=vg_devel1/lv_root rd.lvm.lv=vg_devel1/lv_swap cio_ignore=all,!condev rd.znet=qeth,0.0.0a00,0.0.0a01,0.0.0a02,layer2=1,portno=0'
initial ramdisk...: /boot/initramfs-5.14.0-55.el9.s390x.img component address:
kernel image....: 0x00010000-0x007a21ff
parmline........: 0x00001000-0x000011ff
initial ramdisk.: 0x02000000-0x028f63ff
internal loader.: 0x0000a000-0x0000a3ff
Preparing boot device: dm-0.
Detected SCSI PCBIOS disk layout.
Writing SCSI master boot record.
Syncing disks...
Done.

FCP LUNs that are not part of the root file system, such as data disks, are persistently configured in the file /etc/zfcp.conf. It contains one FCP LUN per line. Each line contains the device bus ID of the FCP adapter, the target WWPN as 16 digit hexadecimal number prefixed with 0x, and the FCP LUN prefixed with 0x and padded with zeroes to the right to have 16 hexadecimal digits, separated by a space or tab.

The WWPN and FCP LUN values are only necessary if the zFCP device is not configured in NPIV mode, when auto LUN scanning is disabled by the zfcp.allow_lun_scan=0 kernel module parameter or when installing RHEL-9.0 or older releases. Otherwise they can be omitted and only the device bus ID is mandatory.

Entries in /etc/zfcp.conf are activated and configured by udev when an FCP adapter is added to the system. At boot time, all FCP adapters visible to the system are added and trigger udev.

Example content of /etc/zfcp.conf:

0.0.fc00 0x5105074308c212e9 0x401040a000000000
0.0.fc00 0x5105074308c212e9 0x401040a100000000
0.0.fc00 0x5105074308c212e9 0x401040a300000000
0.0.fcd0 0x5105074308c2aee9 0x401040a000000000
0.0.fcd0 0x5105074308c2aee9 0x401040a100000000
0.0.fcd0 0x5105074308c2aee9 0x401040a300000000
0.0.4000
0.0.5000

Modifications of /etc/zfcp.conf only become effective after a reboot of the system or after the dynamic addition of a new FCP channel by changing the system’s I/O configuration (for example, a channel is attached under z/VM). Alternatively, you can trigger the activation of a new entry in /etc/zfcp.conf for an FCP adapter which was previously not active, by executing the following commands:

To add a network device that is required to access the root file system, you only have to change the boot options. The boot options can be in a parameter file, however, the /etc/zipl.conf file no longer contains specifications of the boot records. The file that needs to be modified can be located using the following commands:

# machine_id=$(cat /etc/machine-id)
# kernel_version=$(uname -r)
# ls /boot/loader/entries/$machine_id-$kernel_version.conf

Dracut, the mkinitrd successor that provides the functionality in the initramfs that in turn replaces initrd, provides a boot parameter to activate network devices on 64-bit IBM Z early in the boot process: rd.znet=.

As input, this parameter takes a comma-separated list of the NETTYPE (qeth, lcs, ctc), two (lcs, ctc) or three (qeth) device bus IDs, and optional additional parameters consisting of key-value pairs corresponding to network device sysfs attributes. This parameter configures and activates the 64-bit IBM Z network hardware. The configuration of IP addresses and other network specifics works the same as for other platforms. See the dracut documentation for more details.

The cio_ignore commands for the network channels are handled transparently on boot.

Example boot options for a root file system accessed over the network through NFS:

root=10.16.105.196:/nfs/nfs_root cio_ignore=all,!condev rd.znet=qeth,0.0.0a00,0.0.0a01,0.0.0a02,layer2=1,portno=0,portname=OSAPORT ip=10.16.105.197:10.16.105.196:10.16.111.254:255.255.248.0:nfs‑server.subdomain.domain:enc9a0:none rd_NO_LUKS rd_NO_LVM rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYTABLE=us