In the last post we discussed about replacing failed disk in ZFS. Now let us see some of the examples of replacing failed root disk under Solaris volume manager. Note that the failed root disk is under a mirrored configuration.
Replacing the failed root disk
Consider the mirrored root disk configuration as shown below containing the root and swap mirrors and sub mirrors. As shown the disk c1t0d0 is failed causing the root and swap sub mirrors (d11 and d1) to be in maintenance state.
# metastat -c
d10 m 6.0GB d11 (maint) d12 --- swap
d11 s 6.0GB c1t0d0s1 (maint)
d12 s 6.0GB c1t1d0s1
d0 m 74GB d1 (maint) d2 --- root
d1 s 74GB c1t0d0s0 (maint)
d2 s 74GB c1t1d0s0
1. First thing is to remove the sub mirrors in maintenance state. use the metadetach command to detach the failed submirrors :
Syntax :
# metadetach [mirror] [sub-mirror]
Detach the sub-mirrors d11 and d1 :
# metadetach d10 d11 # metadetach d0 d1
2. Next is to remove any metadb on the failed disk. Check for the metadbs configured in the system.
# metadb
flags first blk block count
a u 16 8192 /dev/dsk/c1t0d0s7
a u 8208 8192 /dev/dsk/c1t0d0s7
a u 16400 8192 /dev/dsk/c1t0d0s7
a u 16 8192 /dev/dsk/c1t1d0s7
a u 8208 8192 /dev/dsk/c1t1d0s7
a u 16400 8192 /dev/dsk/c1t1d0s7
As shown above remove the 3 metadbs on c1t0d0s7 slice.
# metadb -d c1t0d0s7
3. We need to remove the disk from the OS before physically removing it from the server. Refer the below post to remove the disk using cfgadm or luxadm.
4. Now add the new disk and run devfsadm to scan for the newly added disk.
# devfsadm
5. Once the disk is detected we need to copy the VTOC table from the c1t0d0 to the newly added disk using the fmthard command :
# prtvtoc /dev/rdsk/c1t0d0s2 | fmthard -s - /dev/rdsk/c1t1d0s2 fmthard: New volume table of contents now in place.
6. Install the bootblk on slice 0 of the new disk.
# installboot /usr/platform/sun4u/lib/fs/ufs/bootblk /dev/rdsk/c1t1d0s0
7. Add the metadbs we deleted.
# metadb -afc 3 c1t0d0s7
8. Update the device ID in the SVM database.
# metadevadm -u c1t0d0
Updating Solaris Volume Manager device relocation information for c1t0d0
Old device reloc information:
id1,sd@n5000cca000368d24
New device reloc information:
id1,sd@n5000c500052fb987
9. Attach the detach sub mirrors using the metaattach command. The syntax to do so is :
Syntax :
# metattach [mirror] [submirror]
This would start the re-sync sub mirrors.
# metattach d10 d11 # metattach d0 d1
10. In case you did not detach the sub mirrors in step 1, run the metereplace command to start the sync which will return the sub mirrors to the ‘Okay’ state:
# metareplace -e d0 c1t0d0s0 d0: device c1t0d0s0 is enabled # metareplace -e d10 c1t0d0s1 d10: device c1t0d0s1 is enabled
11. To check the synching status use metastat -c command :
# metastat -c
d10 m 6.0GB d11 (resync-53%) d12
d11 s 6.0GB c1t0d0s1
d12 s 6.0GB c1t1d0s1
d0 m 74GB d1 (resync-27%) d2
d1 s 74GB c1t0d0s0
d2 s 74GB c1t1d0s0
To continuously check the status of sync every 10 seconds use :
# while true; do metastat | grep done; sleep 10; done
For x86/x64 hardware
In case of a x86/x64 hardware, we need to do some extra steps for the disk replacements. First we need to create a new partition with fdisk before we can copy the VTOC table from the disk c1t1d0.
# fdisk -b /usr/lib/fs/ufs/mboot /dev/rdsk/c1t0d0p0 No fdisk table exists. The default partition for the disk is: a 100% "SOLARIS System" partition Type "y" to accept the default partition, otherwise type "n" to edit the partition table. y
Secondly, we need to install grub instead of bootblk onto the new disk.
# /sbin/installgrub /boot/grub/stage1 /boot/grub/stage2 /dev/rdsk/c1t0d0s0