[Summary] Stress via AdvFS domain...

From: Xavier Mertens (xavier_at_rootshell.be)
Date: 05/22/03

  • Next message: Carl Christopher Bavington _at_ Langley: "Command or shell script to add times?." 
    Date: Thu, 22 May 2003 13:25:50 +0200 (CEST)
To: tru64-unix-managers@ornl.gov

    Its wasn't due to a hardware problem... But one of the disk has their label corrupted. :-(

    I had to reset the box, backup the filesystem, rmfdmn, mkfdmn after a 'disklabel' on the bad disk.

    Xavier 

    --
http://www.rootshell.be
echo '16i[q]sa[ln0=aln100%Pln100/snlbx]sb20293A2058554E494Csnlbxq'|dc
On Wed, 21 May 2003, Xavier Mertens wrote:
> 
> Update...
> 
> The cluster node where the fielset was located crashed and during reboot I got:
> 
> WARNING: cfs_kgs_submit_join_proposal: Join of KCH set FS:ADVFS:NOC_DOMAIN failing with 9d - retrying
> 
> Any advice?
> 
> --
> http://www.rootshell.be
> echo '16i[q]sa[ln0=aln100%Pln100/snlbx]sb20293A2058554E494Csnlbxq'|dc
> 
> On Wed, 21 May 2003, Xavier wrote:
> 
> > 
> > Hi Gurus,
> > 
> > I've a strange behaviour with a fileset in a specific AdvFS domain...
> > 
> > All access to a file in this fileset are blocked (ls, mv, cp, ...) and the processes
> > are left in 'U' state (so un-killable) :/
> > I performed a verfify on the domain but the command blocked in state 'U' too!
> > I can't umount or relocated the fileset to another node...
> > 
> > Any idea?
> > 
> > --
> > http://www.rootshell.be
> > echo '16i[q]sa[ln0=aln100%Pln100/snlbx]sb20293A2058554E494Csnlbxq'|dc
> > 
>
  • Next message: Carl Christopher Bavington _at_ Langley: "Command or shell script to add times?."

    Relevant Pages

    • [Update] Stress via AdvFS domain...
      ... The cluster node where the fielset was located crashed and during reboot I got: ... Any advice? ... On Wed, 21 May 2003, Xavier wrote: ... > I've a strange behaviour with a fileset in a specific AdvFS domain... ...
      (Tru64-UNIX-Managers)
    • Summary: Restoring root fileset on 4.0d
      ... There is a renamefset command that was suggested, but since the system was running (only the clone backups were failing because the root fileset was mistakenly named "roottmp") I did not want to risk renaming the root filesystem in multiuser mode. ... I used the vdump/vrestore process to copy the original data for root, usr, and var to another disk and then I had a tech onsite move the new disk into the scsi position of the original disk. ... I want to restore the root advfs filesystem on a 4.0d system. ...
      (Tru64-UNIX-Managers)
    • Re: Add physical disk resource
      ... 175278 How to add additional non-hot swap SCSI hard disks to a cluster node ... Windows NT/2000/2003 Cluster Technologies ... > I added a new disk that both nodes can see in Windows Disk Management ...
      (microsoft.public.windows.server.clustering)