Andy Cobaugh <[email protected]> writes: > On 2010-03-24 at 13:56, Russ Allbery ( [email protected] ) said:
>> For the rest, that's something that would need to be addressed >> upstream. If you've not already, could you file a bug report with >> [email protected] requesting this feature? It seems like a >> reasonable thing to implement, although so far as I know you're the >> first person who's wanted to do this. > Open a bug with regards to aklog looking in CellServDB directly, or...? Yes, rather than using AFSDB or now SRV records. This may be fixed in 1.5, though; I'm not sure. > I think that makes sense. Would you just rely on AFS_AFSDB in > /etc/openafs/afs.conf.client to tell if we're using afsdb? That would > work in our case. Yes. This is now done in the current packages in unstable. > Though, we're getting our afsd options from OPTIONS in > /etc/openafs/afs.conf, where we've set OPTIONS directly, since we want > to tweak things, and this makes it easy to keep the same afsd options > across debian and redhat, where the openafs.org rpm packages just use > AFSD_OPTIONS in /etc/sysconfig/openafs. You need to ensure the right option is set in /etc/openafs/afs.conf.client and not just command line options set in AFSD_OPTIONS, but afsdb is the default there. > In any case, I think I'd rather have a client that fails to start, or > even crashes, than a client that starts using old db servers after every > package upgrade. Well, from my perspective the correct solution would be to deploy a CellServDB with the correct database servers, but I realize that you don't want to do things that way. -- Russ Allbery ([email protected]) <http://www.eyrie.org/~eagle/> -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
