On Wed, 16 Feb 2011 21:58:07 -0600 "Ryan C. Underwood" <[email protected]> wrote:
> Is it possible to bring the "old release" volume online in the > meantime or am I just screwed until I bring the RW server back and > re-release? I can't find any obvious way to clear the release flag > manually. You may be able to, but depending on when in the release process the release was interrupted, the data in the volume could be very likely to be inconsistent since vos could have been in the middle of restoring to that volume. The way to 'clear' the flag is to remove the site with 'vos remsite' and add it back with 'vos addsite -valid'. Depending on the state of the volume and how long it's been since the interrupted release ocurred, that may or may not actually work. That is, if the fileserver can't attach the volume because the volserver still has it or it looks inconsistent or something, you won't be able to access stuff in the volume no matter what the VLDB says. If that doesn't work, the least error-prone way (and the way easiest to describe) going forward I think is dumping the volume, to make sure you don't lose the data. Once you have a volume dump you can try 'vos restore'ing the dump (possibly under a different name, and then pointing the mountpoint at the new volume), or 'vos convertROtoRW' the RO on the server. If the volserver won't let you get at the volume, you can try "bos restart fs"ing, or just killing the volserver process (bosserver will restart it for you). Just make sure you aren't running any other volume operations when you do that. There's also a voldump utility that can create a volume dump directly from the /vicepX on-disk structure without going through the volserver, which should be a rather safe bet, but I don't think we ship that program. -- Andrew Deason [email protected] _______________________________________________ OpenAFS-info mailing list [email protected] https://lists.openafs.org/mailman/listinfo/openafs-info
