Vgchage is not working

[Sachin Patel]

Hi
On system which has symbois disk array. I reboot the system in single user mode. run command
#vgchange -a y vg01
warnig: array driver is returning an error I/O with errorno 5, on mysystem_001, lun 0, sector 8
vgchange: couldn't attach to the volume group physical volume /dev/dsk/c0t5d0.

What can I do to fix this?
Is it controller problem or disk?

Sachin

[A. Clay Stephenson]

Hi Sachin,

Errno 5 (EIO) is a general I/O error and simply means that a system call was not able to communicate with a device. If you are lucky, it could be something as simple as a loose cable or as bad as two failed array disks (I'm assuming RAID-5). My first shot would be to shutdown the array and the server and carefully check cables and make sure the terminators are in place. If you still have problems, my next attempt would be to try reading using dd another device on the same bus.
e.g. dd if=/dev/rdsk/c0t6d0 of=/dev/null count=1000
If that works the unix scsi controller, cable,
terminators, are ok.

I would then try the same command on the array device. Be VERY careful to only read from the disks and only write to /dev/null.


Hope this gets you started, Clay

[Sachin Patel]

Hi Clay,
I have already reset the controller. change the cable. checked all terminator. I am thinking to hookup vt200 with disk and may be it has more then 2 bad disk and that is why it is not responding. Is it possible?

Sachin

[A. Clay Stephenson]

Yes Sachin,

It is possible. On arrays, if hot spare is not enabled, 2 bad disks will take you down. Three disks must fail to kill you if you have hot spare. It is very important to monitor the health of disk arrays. One of my HP hardware guys told me about a local SA who recently had an array (a Nike) with some funny lights on it.
It turned out that she had 3 bad disks and was toast. It also turned out that she had not tested her backup. She had backed up daily but had never done a test restore and you can guess the rest. She was asked not to return to work the next day. I tell you this not to scare you but to drive home the importance of running and testing backups.

In you case, since it is reporting a specific sector, it appears than an lseek() failed and that would most likely be a corrupt LUN due to multiple failed disks. It could also be the array controller itself. In any event, I would hook the terminal up and hopefully you have a rather simple problem rather than the horror show.

Best of luck, Clay

[A. Clay Stephenson]

Hi Sachin,

One other thought is that you just might have a corrupt /etc/lmvtab. That would be nice at this point. If your array passes its diagnostics, I would then do the dd test. If that is ok, then it is very possible that the lvmtab is corrupt and that a vgscan could restore you to health.

Clay

[Shahul]

Hi

Just try this.

#cd /etc
#mv lvmtab lvmtab.old
#vgscan
#vgchange -a y vgxx

Now checkup whether it is giving any error or not.

Best of luck
Shahul

[Sachin Patel]

Hi
I am going up there now. I will let you know what was wrong.

Sachin

[Sachin Patel]

I hookup the VT220 with symbois disk controller
It is asking me for passowrd to login. Is there a default password for all this controller?
(may be simpicity or symc or.....)
Any one?

Sachin

[Sachin Patel]

Finally found the password it is standard
"infiniti"

Sachin