Clearing stale mirrored disks

[Dennis Weitzell]

I have searched the forums and manuals but have not been able to find a clear answer to a question I have. We have a server (R390 running HP-UX B.11.00) that has several mirrored disks on a LV that are stale (see below).

/usr/sbin/lvdisplay -v /dev/vg00/lvol1
--- Logical volumes ---
LV Name /dev/vg00/lvol1
VG Name /dev/vg00
LV Permission read/write
LV Status available/stale
Mirror copies 1
Consistency Recovery MWC
Schedule parallel
LV Size (Mbytes) 84
Current LE 21
Allocated PE 42
Stripes 0
Stripe Size (Kbytes) 0
Bad block off
Allocation strict/contiguous
IO Timeout (Seconds) default

--- Distribution of logical volume ---
PV Name LE on PV PE on PV
/dev/dsk/c1t6d0 21 21
/dev/dsk/c1t4d0 21 21

--- Logical extents ---
LE PV1 PE1 Status 1 PV2 PE2 Status 2
0000 /dev/dsk/c1t6d0 0000 stale /dev/dsk/c1t4d0 0000 current
0001 /dev/dsk/c1t6d0 0001 stale /dev/dsk/c1t4d0 0001 current
0002 /dev/dsk/c1t6d0 0002 stale /dev/dsk/c1t4d0 0002 current
0003 /dev/dsk/c1t6d0 0003 stale /dev/dsk/c1t4d0 0003 current
0004 /dev/dsk/c1t6d0 0004 stale /dev/dsk/c1t4d0 0004 current
0005 /dev/dsk/c1t6d0 0005 current /dev/dsk/c1t4d0 0005 current
0006 /dev/dsk/c1t6d0 0006 stale /dev/dsk/c1t4d0 0006 current
0007 /dev/dsk/c1t6d0 0007 stale /dev/dsk/c1t4d0 0007 current
0008 /dev/dsk/c1t6d0 0008 stale /dev/dsk/c1t4d0 0008 current
0009 /dev/dsk/c1t6d0 0009 current /dev/dsk/c1t4d0 0009 current
0010 /dev/dsk/c1t6d0 0010 current /dev/dsk/c1t4d0 0010 current
0011 /dev/dsk/c1t6d0 0011 current /dev/dsk/c1t4d0 0011 current
0012 /dev/dsk/c1t6d0 0012 current /dev/dsk/c1t4d0 0012 current
0013 /dev/dsk/c1t6d0 0013 current /dev/dsk/c1t4d0 0013 current
0014 /dev/dsk/c1t6d0 0014 current /dev/dsk/c1t4d0 0014 current
0015 /dev/dsk/c1t6d0 0015 current /dev/dsk/c1t4d0 0015 current
0016 /dev/dsk/c1t6d0 0016 current /dev/dsk/c1t4d0 0016 current
0017 /dev/dsk/c1t6d0 0017 current /dev/dsk/c1t4d0 0017 current
0018 /dev/dsk/c1t6d0 0018 current /dev/dsk/c1t4d0 0018 current
0019 /dev/dsk/c1t6d0 0019 current /dev/dsk/c1t4d0 0019 current
0020 /dev/dsk/c1t6d0 0020 current /dev/dsk/c1t4d0 0020 current

This is a production server running Oracle. The questions I have are:
1) Can I run vgsync with users online without causing a problem or would it be better to run this after hours?
2) Also, will it hurt to have Oracle running or should it be stopped before resyncing the disks?
3) Will running vgsync take long?

Any feedback will be appreciated.
Thanks,
Dennis

[A. Clay Stephenson · Accepted solution]

It is perfectly safe to run vgsync or lvsync while the system is operational. Oracle will be fine. The length of time varies with the number of stale extents but it probably limited to a few tens of seconds. Full vgsync's (e.g. when replacing a disk 'on the fly') can take tens of minutes for large, busy disks.

You really should be concentrating on why you are seeing these stale extents. You probably have a disk that is going to fail completely soon.

[Dennis Weitzell]

Thanks for the quick reply.

We have had a script running on all our servers for some time that sends the admins an email if a "stale" disk is seen. Since this problem just showed up on this server, what would be the best way to determine which disk is having problems & when to replace it?

[Dennis Weitzell]

Duh. Sorry, I can see which disk is having the problem.