stale extents on mirrored disks on different sectors

[A.K.]

Hi Guys,
I have a server running HPUX 11.00
Two internal root disks mirrored(primary and alternate) with mirror/ux
Both disks have stale extents but on different sectors
--- Logical extents ---
LE PV1 PE1 Status 1 PV2 PE2 Status 2
00000 /dev/dsk/c1t15d0 00277 current /dev/dsk/c3t15d0 02847 current
00001 /dev/dsk/c1t15d0 00278 current /dev/dsk/c3t15d0 02848 current
00002 /dev/dsk/c1t15d0 00279 current /dev/dsk/c3t15d0 02849 current
00003 /dev/dsk/c1t15d0 00280 current /dev/dsk/c3t15d0 02850 current
00004 /dev/dsk/c1t15d0 00281 current /dev/dsk/c3t15d0 02851 current
00005 /dev/dsk/c1t15d0 00282 current /dev/dsk/c3t15d0 02852 current
00006 /dev/dsk/c1t15d0 00283 current /dev/dsk/c3t15d0 02853 stale
00007 /dev/dsk/c1t15d0 00284 stale /dev/dsk/c3t15d0 02854 current
00008 /dev/dsk/c1t15d0 00285 current /dev/dsk/c3t15d0 02855 current
00009 /dev/dsk/c1t15d0 00286 current /dev/dsk/c3t15d0 02856 current
00010 /dev/dsk/c1t15d0 00287 current /dev/dsk/c3t15d0 02857 current

I have this issue on lvol1(boot) , lvol3 (root) and lvol7 (/var)
How can I fix this problem?

thanks,
A.K.

[Khairy]

hi a.k

you could sync you the affected LV.

eg:
# lvsync /dev/vg00/lvol1

If it fails, then check for disk status.

Check output from the following and look for 'lbolt scsi'.

# dmesg
# more /var/adm/syslog/syslog.log

To confirm, run the following (in onlinediag is installed)
# echo "map selall information;wait infolog quit ok" | cstm > /tmp/cstm.txt

View the file and check your vg00 disk. Look for read/write errors. If the value is more than 0 and lbolt scsi errors occured in syslog or dmesg. This indicate disk problem.

Action: Get make_tape_recovery backup ready and replace the disks!

really appreciate if you could assign some points if this answer yr question, thanks!

[Michael Steele_2]

Minor problem. As pointed out above 'vgsync' to correct the problem. Post results please.

With soooo few stale extents (* two ? *) I wouldn't call this a disk problem yet.

[Mridul Shrivastava]

Just to add if this is the status on other LVs of this VG then you could try vgsync as well so it will sync all the LVs.

Remember these commands may take long time depending upon the size of LVs.

[A.K.]

Well,

dmesg shows the following
LVM: Failed to automatically resync PV 1f01f000 error: 5
LVM: Failed to automatically resync PV 1f01f000 error: 5
LVM: Failed to automatically resync PV 1f01f000 error: 5
LVM: Failed to automatically resync PV 1f01f000 error: 5
LVM: Failed to automatically resync PV 1f01f000 error: 5

lvsync shows the following

lvsync /dev/vg00/lvol7
lvsync: Couldn't re-synchronize stale partitions of the logical volume:
I/O error
lvsync: Couldn't resynchronize logical volume "/dev/vg00/lvol7".



pvdisplay shows the following

pvdisplay /dev/dsk/c3t15d0
--- Physical volumes ---
PV Name /dev/dsk/c3t15d0
VG Name /dev/vg00
PV Status available
Allocatable yes
VGDA 2
Cur LV 22
PE Size (Mbytes) 4
Total PE 4340
Free PE 83
Allocated PE 4257
Stale PE 3
IO Timeout (Seconds) default
Autoswitch On

pvdisplay /dev/dsk/c1t15d0
--- Physical volumes ---
PV Name /dev/dsk/c1t15d0
VG Name /dev/vg00
PV Status available
Allocatable yes
VGDA 2
Cur LV 22
PE Size (Mbytes) 4
Total PE 4340
Free PE 83
Allocated PE 4257
Stale PE 4
IO Timeout (Seconds) default
Autoswitch On


The following command
echo "map selall information;wait infolog quit ok" | cstm > /tmp/cstm.txt

Does not produce any information about vg00



I also run dd command as follow and now it hangs and I can not kill it

dd if=/dev/rdsk/c1t15d0 of=/dev/null bs=1024k count=64

Does not seems as minor problem to me â ¦.:c(
what should I do ?

Thanks ,
A.K.

[sachit patil]

Hi
DD command is very dangerous command.
If DD is not working on the disk, then it might have problem with disk.
You can try diskinfo and pvdisplay -v vgname to get no of stales on disk.


DD will not be killed . For that you have to reboot the serbver.


Regards
Sachit

[sachit patil]

Best solution is replace the disk.

[Siju Vadakkan]

Since lvsync and dd are failed. The only ways is to replace the disk.

if it hot swappable replace it online.
You can execute the following steps.
remove the disk from the server and insert the new one in the same slot.

#ioscan -fnC disk
#insf -eC disk
#vgcfgrestore -n vg00 /dev/rdsk/c3t15d0
#mkboot -b /dev/rdsk/c1t15d0 /dev/rdsk/c3t15d0
#vgsync vg00
#lvlnboot -R

[Michael Steele_2]

If 'dd' hangs then its a bad disk. If the disk is hot swappable then you can sometimes kill the 'dd' command if you pull out the disk and then replace it.

Examine the disk, a red led should be lit up if its failed.

You do have a bad disk. It does need to be replaced.