LV missing after server's unexpected shutdown

[Maxim Yakimenko]

Hi, collegues

I have following trouble:

Server halted due to machine check, after reboot I try to activate volume group with Oracle data on it - it have ~240 LVs.

I do "vgimport -m vgdb1.map -f vgdb1.pv vgdb1"

For some LVs vgimport complains:

Logical Volume is not defined on any physical volume.
"/dev/vgdb1/lv259" is missing Physical Volumes.


I checked PVs - it still correct, after reboot pathes and DSF names are still unchanged.

Data on theese PVs appears by synchronization via HP XP CA, it seems that machine shutdown has led to crash of internal LVM structures on PVs. I wonder - can I recover from this failure or I have to resync devices? For some reason I dont want to do this.

[Michal Kapalka (mikap)]

hi,

check the PV with pvdisplay.

mikap

[Maxim Yakimenko]

Well, I did pvdisplay -v on first PV of group and found that info about some LVs extents located on this PV is gone. What's next? As I understand I have only one option - resync?

[Michal Kapalka (mikap)]

hi,

normally, you should check :

1. chcek MP ( vfp / event log )
2. /etc/rc.log
....

3. dmesg
2. ioscan -fn |grep NO_HW
3. grep EMS /var/adm/syslog/syslog.log


if you have som problem with the VG...

pvdisplay ( diskinfo ... )

- what is your disk configuration ??

check the disk on the Storage for some errors, you could also use xpinfo to gather more information.

....


mikap

[kashyapi]

you can try running vgchid on the PVs
Then try to import using following command..
#vgimport -v -s -m VGName

you can run this command only if you had taken map file using -s option.