Thanks to Jason Orendorf, Pat O'Brien, Maria Gililand and Oisin McGuiness
(did I miss anybody? apologies if I did) for your helpful replies. All of
you were right.
I did have a problem with one of my disks. The operator on duty wrongly
noted to me that no fault-lights were showing, actually there were.
Comparing the output of ls -l /dev/rz* with 'scu show edt' confirmed that a
disk was indeed out of action. I went to the location of that server (it is
offsite), fiddled with the disk and with LSM to find that the O/S would not
read the disk (i/o error every time), and have logged a call to have the
disk replaced.
Now awaiting the arrival of a fresh disk. Hope to be able to quickly find
the right LSM commands to put things back to normal after that.
> -----Original Message-----
> From: John Tan
> Sent: Tuesday, July 03, 2001 4:00 PM
> To: tru64-unix-managers_at_ornl.gov
> Subject: Unstartable LSM plex
>
> Hi. I have researched a problem to this point, but am now needing help to
> continue.
>
> Problem started with difficulty logging into a server. I found the
> problem to be because of an Advfs issue. We run Advfs on LSM volumes in
> this case.
>
> According to /var/adm/syslog.dated/kernel.log:
> ul 3 14:32:52 asx090 vmunix: AdvFS I/O error:
> Jul 3 14:32:52 asx090 vmunix: Domain#Fileset: bcpfas_dom#d1
> Jul 3 14:32:52 asx090 vmunix: Mounted on: /d1
> Jul 3 14:32:52 asx090 vmunix: Volume: /dev/vol/movedg/mvdgvol1
> Jul 3 14:32:52 asx090 vmunix: Tag: 0x00000001.8001
> Jul 3 14:32:52 asx090 vmunix: Page: 2408
> Jul 3 14:32:53 asx090 vmunix: Block: 42214304
> Jul 3 14:32:53 asx090 vmunix: Block count: 128
> Jul 3 14:32:53 asx090 vmunix: Type of operation: Read
> Jul 3 14:32:53 asx090 vmunix: Error: 5
> Jul 3 14:32:53 asx090 vmunix: io/vol.c(volerror): Uncorrectable read
> error on v
> olume mvdgvol1, plex mvdgvol1p1, block 43001888
>
> This problem keeps reoccuring.
>
> To get over the problem, I initially rebooted since the system was too
> slow to be able to do any work. System is now performing okay, but I am
> no longer able to mount the LSM volume.
>
> mount says:
> bcpfas_dom#d1 on /d1: Bad file number
> bcpfas_dom#d2 on /d2: Bad file number
> bcpfas_dom#d3 on /d3: Bad file number
> bcpfas_dom#d4 on /d4: Bad file number
> bcpfas_dom#d5 on /d5: Bad file number
> bcpfas_dom#d6 on /d6: Bad file number
> bcpfas_dom#d17 on /d17: Bad file number
>
> /sbin/advfs/verify says:
> verify: can't get set info for domain 'bcpfas_dom'
> verify: error = Bad file number
> +++ Domain verification +++
>
> main: unable to get info for domain 'bcpfas_dom'
> error: 9, Bad file number
>
> I search /etc/fdmns to note that domain bcpfas_dom is built out of a LSM
> diskgroup called "movedg", which contains one volume "movedgvol1", and one
> plex within that volume "movedgvol1p1".
>
> volume -g movedg startall says:
> gen/volume: Volume mvdgvol1 has no CLEAN or non-volatile ACTIVE plexes
>
> volmend fix says that no plexes are in stale state.
>
> volinfo -pg movedg says:
> vol mvdgvol1 gen Unstartable
> plex mvdgvol1p1 NODEVICE
>
> Therefore, please help. How go I get the plex recognisable as a device,
> and the volume startable?
>
> Your appreciation is greatly appreciated. Thanks.
>
Received on Wed Jul 04 2001 - 04:42:37 NZST