AIX - stale partition

UNIX AIX
1

Hi everybody,
I have a little problem with my AIX 6.1, PowerHA 6.1 LVM mirror. After problem with SAN pathing of our one Datacenter, I have LV at stale state.

 # lsvg cpsdata2vg
VOLUME GROUP:       cpsdata2vg               VG IDENTIFIER:  00fb518c00004c0000000169445f4c2c
VG STATE:           active                   PP SIZE:        1024 megabyte(s)
VG PERMISSION:      read/write               TOTAL PPs:      6142 (6289408 megabytes)
MAX LVs:            256                      FREE PPs:       441 (451584 megabytes)
LVs:                2                        USED PPs:       5701 (5837824 megabytes)
OPEN LVs:           2                        QUORUM:         2 (Enabled)
TOTAL PVs:          2                        VG DESCRIPTORS: 3
STALE PVs:          1                        STALE PPs:      108
ACTIVE PVs:         1                        AUTO ON:        no
Concurrent:         Enhanced-Capable         Auto-Concurrent: Disabled
VG Mode:            Concurrent
Node ID:            1                        Active Nodes:       2 3 4
MAX PPs per VG:     32768                    MAX PVs:        1024
LTG size (Dynamic): 256 kilobyte(s)          AUTO SYNC:      no
HOT SPARE:          no                       BB POLICY:      relocatable
MIRROR POOL STRICT: off
PV RESTRICTION:     none                     INFINITE RETRY: no
DISK BLOCK SIZE:    512                      CRITICAL VG:    no

 # lspv
...
hdisk36         00fb518c4457e71a                    cpsdata2vg      concurrent
hdisk37         00fb518c4457f895                    cpsdata2vg      concurrent

 # lsvg -l cpsdata2vg
cpsdata2vg:
LV NAME             TYPE       LPs     PPs     PVs  LV STATE      MOUNT POINT
cpsabcd2lv          jfs2       2850    5700    2    open/stale    /cpsabcd2
loglv00             jfs2log    1       1       1    open/syncd    N/A

 # lsvg -p cpsdata2vg
cpsdata2vg:
PV_NAME           PV STATE          TOTAL PPs   FREE PPs    FREE DISTRIBUTION
hdisk36           active            3071        220         00..00..00..00..220
hdisk37           missing           3071        221         00..01..00..00..220

Normally I solve it by command varyonvg for non-concurrent LUNs. I cant find solutions for Enhanced-Capable concurrent LUNs.
I guess that the right way to solve this problem is command varyonvg too. With parameter '-c' for concurrent vary.

Does anyone have experience with this procedure? I cant umount this filesystem right now, so I am curious if I can do this without unavailability or other problem with filesystem.

Thank you.

2

You have lost access to a disk, hdisk37

How is this provided? If it's a real disk, then it's broken. If it's SAN provided, then something has gone wrong in the provision.

If you can get the disk back online, you might get away with syncvg -l cpsabcd2lv to re-mirror, but you need the hardware (however provided) to be sorted first.

I notice that loglv00 is not mirrored. I would suggest getting that done. If you had lost hdisk36 then would be in real trouble. As it is, LVM is coping.

Kind regards,
Robin

3

Thank you for advice.

It´s SAN provided. Each disk in mirror is from other DC/path. We lost path to one of the datacenters. It´s already repaired.

I know that

syncvg

is called by

varyonvg

, which I´m used to use on stale partition. Is there significant difference between using these two commands?

I must definitely resolve loglv00 as soon as possible.

4

My apologies for being away for a few days. There should be no difference with the process because they will both do the same thing. If the lost LUN is now available and visible to the volume group, then the sync, however it runs will have to read the appropriate physical partitions from the good side and copy them to the stale side if the PP has been updated. It will not copy the data if the PP has not updated, but for a large LV, this process can take some time to complete.

You should be able to monitor it from another session so you get an idea of progress and see the stale PP count reducing.

Robin

5

I already repaired it. I ran the command

I saw the

in the processes while repairing.

# lsvg -l cpsdata2vg
cpsdata2vg:
LV NAME             TYPE       LPs     PPs     PVs  LV STATE      MOUNT POINT
cpsabcd2lv          jfs2       2850    5700    2    open/syncd    /cpsabcd2
loglv00             jfs2log    1       1       1    open/syncd    N/A

It is OK now.

Thank you very much.

6

I would suggest mirroring your loglv00 now too.

Kind regards,
Robin

8

Hi,

I already made mirroring for the loglv00.

mklvcopy -e x loglv00 2 hdisk37

The parameter -e was probably a little extra here.

I tested this on test environment, if I dont need mapfile here. It always filled the same (free) PP on opposite disk.

After that I again ran command:

varyonvg -c cpsdata2vg

loglv00 is mirrored now.

# lsvg -l cpsdata2vg
cpsdata2vg:
LV NAME             TYPE       LPs     PPs     PVs  LV STATE      MOUNT POINT
cpsabcd2lv          jfs2       2850    5700    2    open/syncd    /cpsabcd2
loglv00             jfs2log    1       2       2    open/syncd    N/A

# lslv -m loglv00
loglv00:N/A
LP    PP1  PV1               PP2  PV2               PP3  PV3
0001  0617 hdisk36           0617 hdisk37

Hope that this will help somebody.

Have a nice day,
Andrew.