FDR/Upstream for zLinux disaster recovery using the SLES10 Installer

From LinuxVM
Jump to: navigation, search

FDR/Upstream for zLinux disaster recovery using the SLES10 installer

Assumptions: 1. You have created a Linux guest with all the right disk drives attached at all the right addresses. In this case we have a mod 9 boot volume at 100, swap volumes at 102, 103, 104, 105, 106, 107, and a data volume at 200. 2. You have booted the Linux guest to a SSH commands prompt. 3. You have placed the usdr.tar file on an accessible NFS server. 4. You know your lvm configuration. That is, you physical volumes, your volume groups, your logical volumes, and what filesystems are mounted on what logical volumes.

Notes: 1. We have some reiserfs file systems, so we have to load that module. 2. This procedure has it's roots in a procedure that was developed for SLES9. It has not been fully updated for SLES10. 3. This version of the procedure uses lvm for all filesystems except for /boot. Our new clone base uses a different lvm layout, I have not updated the examples in this document.

This procedure works with 3.5.0.f of the Upstream client.


Steps:

1. Load the DASD device driver.

  modprobe dasd_eckd_mod

2. Make all the dasd devices available

  dasd_configure 0.0.0100 1 0
  dasd_configure 0.0.0102 1 0
  dasd_configure 0.0.0103 1 0
  dasd_configure 0.0.0104 1 0
  dasd_configure 0.0.0105 1 0
  dasd_configure 0.0.0106 1 0
  dasd_configure 0.0.0107 1 0
  dasd_configure 0.0.0200 1 0

3. Verify that all devices came online

cat /proc/dasd/devices
0.0.0100(ECKD) at ( 94:  0) is dasda  : active at blocksize 4096, 1802880 blocks, 7042 MB
0.0.0102(ECKD) at ( 94:  4) is dasdb  : active at blocksize 4096, 600840 blocks, 2347 MB
0.0.0103(ECKD) at ( 94:  8) is dasdc  : active at blocksize 4096, 600840 blocks, 2347 MB
0.0.0104(ECKD) at ( 94: 12) is dasdd  : active at blocksize 4096, 600840 blocks, 2347 MB
0.0.0105(ECKD) at ( 94: 16) is dasde  : active at blocksize 4096, 600840 blocks, 2347 MB
0.0.0106(ECKD) at ( 94: 20) is dasdf  : active at blocksize 4096, 600840 blocks, 2347 MB
0.0.0107(ECKD) at ( 94: 24) is dasdg  : active at blocksize 4096, 600840 blocks, 2347 MB
0.0.0200(ECKD) at ( 94: 28) is dasdh  : active at blocksize 4096, 1802880 blocks, 7042 MB

4. Make a cross-reference of the device addresses and device names for later use. For example 0.0.0100 is dasda.

5. Load the kernel modules that allow the use of lvm.

  modprobe dm-mod
  modprobe reiserfs
  /sbin/devmap_mknod.sh

6. Format the drives for Linux use.

  dasdfmt –b 4096 –d cdl –p –f /dev/dasda
  Repeat for each dasd device found above.

7. Partition the boot volume.

  fdasd /dev/dasda
  add  partition

n

  first track

2

  last track

2114

  add a partition

n

  first track
	2115
  last track  (for a mod 9, use 150239)

150239

  write the partition table and exit

w 8. Partition the remaining volumes

  fdasd -k -a /dev/dasdb
  (Repeat for all the remaining dasd devices)

9. Create the physical volumes

  pvcreate /dev/dasda2
  pvcreate /dev/dasdc1

10.Create the volume groups.

  vgcreate  SapVG  /dev/dasdc1 
  vgcreate  System  /dev/dasda2

11.Activate the volume groups

  vgscan
  vgchange -ay

12.Create the logical volumes.

  lvcreate -L 100.00M -n CTD SapVG
  lvcreate -L 152.00M -n CTM SapVG
  lvcreate -L 4.00M -n SapLV SapVG
  lvcreate -L 4.00M -n SapLV1 SapVG
  lvcreate -L 4.00M -n SapLV2 SapVG
  lvcreate -L 4.00M -n SapLV3 SapVG
  lvcreate -L 4.00M -n SapLV4 SapVG
  lvcreate -L 700.00M -n SapLV5 SapVG
  lvcreate -L 1.00G -n SapLV6 SapVG
  lvcreate -L 4.00G -n SapLV7 SapVG
  lvcreate -L 128.00M -n home System
  lvcreate -L 1.00G -n opt System
  lvcreate -L 512.00M -n root System
  lvcreate -L 92.00M -n srv System
  lvcreate -L 164.00M -n tmp System
  lvcreate -L 2.00G -n usr System
  lvcreate -L 1.00G -n var System

13.Make the swap spaces.

  mkswap  /dev/dasdb1
  mkswap  /dev/dasdc1
  mkswap  /dev/dasdd1
  mkswap  /dev/dasdf1
  mkswap  /dev/dasdg1
  

14.Make the filesystems.

  mkreiserfs -b 4096 -q /dev/system/root
  mke2fs -b 4096 -q /dev/dasda1
  mkreiserfs -b 4096 -q /dev/System/home
  mkreiserfs -b 4096 -q /dev/System/opt
  mkreiserfs -b 4096 -q /dev/System/srv
  mkreiserfs -b 4096 -q /dev/System/tmp
  mkreiserfs -b 4096 -q /dev/System/usr
  mkreiserfs -b 4096 -q /dev/System/var
  mkreiserfs -b 4096 -q /dev/SapVG/CTD
  mkreiserfs -b 4096 -q /dev/SapVG/CTM
  mke2fs -b 4096 -q /dev/SapVG/SapLV1
  mke2fs -b 4096 -q /dev/SapVG/SapLV4
  mkreiserfs -b 4096 -q /dev/SapVG/SapLV6
  mke2fs -b 4096 -q /dev/SapVG/SapLV2
  mkreiserfs -b 4096 -q /dev/SapVG/SapLV5
  mkreiserfs -b 4096 -q /dev/SapVG/SapLV7
  mke2fs -b 4096 -q /dev/SapVG/SapLV3

15.Make the mount points and mount the filesystems.

  mkdir -p /mnt/
  mount -t reiserfs /dev/system/root /mnt/
  mkdir -p /mnt/boot
  mount -t ext2 /dev/dasda1 /mnt/boot
  mkdir -p /mnt/home
  mount -t reiserfs /dev/System/home /mnt/home
  mkdir -p /mnt/opt
  mount -t reiserfs /dev/System/opt /mnt/opt
  mkdir -p /mnt/srv
  mount -t reiserfs /dev/System/srv /mnt/srv
  mkdir -p /mnt/tmp
  mount -t reiserfs /dev/System/tmp /mnt/tmp
  mkdir -p /mnt/usr
  mount -t reiserfs /dev/System/usr /mnt/usr
  mkdir -p /mnt/var
  mount -t reiserfs /dev/System/var /mnt/var
  mkdir -p /mnt/controld
  mount -t reiserfs /dev/SapVG/CTD /mnt/controld
  mkdir -p /mnt/controlm
  mount -t reiserfs /dev/SapVG/CTM /mnt/controlm
  mkdir -p /mnt/idoc
  mount -t ext2 /dev/SapVG/SapLV1 /mnt/idoc
  mkdir -p /mnt/sapmnt
  mount -t ext2 /dev/SapVG/SapLV4 /mnt/sapmnt
  mkdir -p /mnt/sapmnt/PRD
  mount -t reiserfs /dev/SapVG/SapLV6 /mnt/sapmnt/P10
  mkdir -p /mnt/usr/sap
  mount -t ext2 /dev/SapVG/SapLV2 /mnt/usr/sap
  mkdir -p /mnt/usr/sap/PRD
  mount -t reiserfs /dev/SapVG/SapLV5 /mnt/usr/sap/P10
  mkdir -p /mnt/usr/sap/sort
  mount -t reiserfs /dev/SapVG/SapLV7 /mnt/usr/sap/sort
  mkdir -p /mnt/usr/workdata
  mount -t ext2 /dev/SapVG/SapLV3 /mnt/usr/workdata

16. Create and mount the FDR/Upstream client directory.

   mke2fs –vm0 /dev/ram3
   mount –t ext2 /dev/ram3 /tmp
   mkdir /tmp/fdrupstream

17. Mount the directory containing the usdr.tar file

   mount -t nfs 10.10.10.10:/zlinux /mnt

18. Extract the FDR/UPstream Disaster Recovery image.

   cd /tmp/fdrupstream
   tar -xvf /mnt/usdr.tar
   umount /mnt

19. Configure Upstream to communicate with the Storage Server.

   ./uscfg 
   (Contact your FDR/Upstream administrator for IP addresses and ports that need to be entered in the configuration tool.)

20. Start the recovery utility.

   Obtain the following information.
   The FDR/Upstream profile you are going to restore.
   The RACF Userid and Password.
   ./usdr
   Enter the requested information.
   Profile name
   Userid
   Password
   Hit Enter
   Select appropriate remote version
   Select Highlighted back to full
   Select Allow any file system type
   Select Begin Restore
   Select Yes to Are you ready to begin the system restore.

21. The FDR restore does not perform zipl to make the restored system bootable. Issue the following commands.

   chroot /mnt

22. Using your favorite editor verify that /etc/zipl.conf is set to appropriate values. Then run the zipl command.

   zipl -V

23. Use the following commands to back out gracefully.

   exit
   cd /
   umount -a
   vgchange -a n
   

24. The following commands back gracefully out of the installer.

   yast
   Tab to the Abort option and press enter
   Tab to the Abort Installation option and press enter.
   (Your SSH connection should go away.)
   

25. The following shuts down the installer kernel.

   Select option 8 to power off
   Select option 1 yes.

26. IPL the restored system. Please be aware that unless you have defined your logical volumes in the same sequence as they were originally defined, some of you file systems may not contain what you think they ought to. The following commands should remedy the situation.

If the root file system is not mounted rw. Try this command.

  mount –n –oremount,rw

The following command will rebuild the lvm nodes.

  vgscan --mknodes

While you are at it, it may be helpful to rewrite the boot area

  mount –t ext2 /dev/dasda1 /boot
  zipl 

Reboot the system and look for errors.

If you are having a really bad day, it may be necessary to run the vgscan --mknodes command under the SLES10 installer. (I have had to do this under SLES9.) Boot the installer. Load the device driver. Step 1. Make the drives available. Step 2. Load the kernel modules that allow the use of lvm. Step 5. Activate the volume groups. Step 11. Make the mount points and mount the filesystems. Step 15. chroot /mnt vgscan --mknodes Using your favorite editor verify that /etc/zipl.conf is set to appropriate values. Then run the zipl command. Step 22. Use the following commands to back out gracefully. Step 23. The following commands back gracefully out of the installer. Step 24. Boot the restored system and check for errors.

27. This procedure was developed from an email from the FDR/Upsteam folks way back in the early SLES9 days. Information for updating it work with SLES10 came from a SHARE presentation "Linux on z/VM System Programmer Survival Guide" by Jay Brenneman.