problem on old K200

[Clint Gibler]

I have a legacy system that is running HPUX 10.20 a hard drive appears to have gone bad on one of the has units, so now when it boots up my vg10 isn't getting loaded vg10 consisted of 2 discs one was a mirror and it appears that one is ok but when I go into sam it tells me that the 1 disc is either not there anymore or can't be read. it is claimed under ioscan but when I attempt to activate the volume group vg10 it just says it can't connect the one disc I have even tried vgchange -a y -q n vg10 for the quorum whenever I do vgchange I get the error and then the server reboots itself for some reason. I need to get the server to load vg10 just using the mirror disk somehow until I can get a replacement for the bad disk.

[Steven E. Protter]

Shalom Clint,

You can not activate the volume group because quorum is not present, which is more than 50% of the members up. The risk of a 2 disk vg.

You can vgexport the volume group and then rebuild it with vgcreate and 1 disk to bring th e one working disk online.

SEP

[Clint Gibler]

Steven so would that keep all of the data intact on that disk? If quorom is not present how come trying to activate with quorom set to no deosn't work?

[Mridul Shrivastava]

Quorum is checked if we are not using -q n so If you execute vgchange with -q n option it should not check for quorum and activate the vg.
How its rebooting the server could you please post the error when you execute this command and how come it reboots the server. Eventhough if this command is not executed successfully It should not reboot the server.
Please post error message and check OLDsyslog.log and syslog.log for any errors related to this.

[OldSchool]

I'm not sure, but I *believe*

vgreduce -f vg10

will remove the dead disk from the volume group definition. you should then be able to activate it.

[Albert_31]

Hello Clint,

wierd issue i say..looks like someone played with the lvm info or some bad h/w .. can you post the following so we can see what is the current situation

# vgcfgrestore -l -f /etc/lvmconf/vg10.conf
# vgcfgrestore -l -f /etc/lvmconf/vg10.conf.old
# strings /etc/lvmtab
collect the vgid and pvid
#xd -j8200 -N16 /dev/rdsk/c*t*d* (replace * as required)
#uname -i

To remove the bad mirror copy, we have to reduce the Lvol's first and for that..the vg should be activated which is not in ur case.. one more way is to remove the disk physically and then try to activate the vg..

Try this command
# vgchange -a y -q n -s /dev/vg10

If the system is panicking then it has to be either h/w related or a patch issue..

regards

albert