2. if the second (old crashed) server comes up , the volume on this server will not switched to online when the fileserver starts up ..
Do all parties involved (fileserver hosting the converted RO volume, crashed fileserver, vlserver) need to be openafs-1.3? What if one of them isn't?
The recovery process Sven describes is somewhat manual. Basically, you're converting one of the RO copies to a RW, and then telling the VLDB of the new location of the RW volume. If the crashed fileserver comes back, it will have an old copy of the RW volume online, which you'll want to get rid of with 'vos zap', but the vldb will be pointing to the new copy, so it won't really matter much. The key thing to bear in mind here is that the VLDB never gets information about volume locations from fileservers; that information changes only when updated by some administrative tool.
To convert an RO volume to RW, you need a version of vos that has the convertrotorw command, and a fileserver new enough to support the AFSVolConvertROtoRWVolume RPC. There is no requirement on the version of the crashed fileserver (it's not involved), or of the VLDB.
A similar question: Can I "vos restore" a volume dump larger than 2 GB (without large files, from a 1.2 or 1.3 server) onto a 1.2 fileserver?
Yes.
-- Jeffrey T. Hutzelman (N3NHS) <[EMAIL PROTECTED]> Sr. Research Systems Programmer School of Computer Science - Research Computing Facility Carnegie Mellon University - Pittsburgh, PA
_______________________________________________ OpenAFS-info mailing list [email protected] https://lists.openafs.org/mailman/listinfo/openafs-info
