[OpenAFS] OpenAFS with RAID

[Esther Filderman]

> I believe this scenario will not work because the VLDB entries for all
> of the volumes that are being mounted by Server B are listed as being
> on Server A.  Since Server A is unreachable, the volume server when
> performing the "vos syncvldb" and "vos syncserv" steps will not be able
> to verify that the volume is no longer on Server A.  Therefore, there
> is now an "irreconcilable conflict" that will cause the vos command to
> "write a message to the standard error stream."  The vos "command never
> removes volumes from file server machines."  The quotes are from the man
> pages for "vos syncvldb" and "vos sycnserv".

Having run into just this situation a few times lately, I can speak of
this situation. It's a time consuming pain in the neck to get out of.

The short form is:

You need to identify every RW volume believed by the vldb to be on
that machine ("vos listvldb -server A -partition {a,b,c, etc}" and
tell the vldb it lives somewhere else.

The easiest way to to do this is to use "vos changeloc" to "move" it
to any partition on any responding fileserver, then use "vos syncvldb"
on said fileserver to properly re-sync the vldb.

If you're one of those people who writes perl scripts to change the TP
in your bathroom, you  can probably whip something up to "changeloc"
to the correct place, via a "listvdb" of the dead server and a
"listvol" of living servers, but I would still be doing a syncvldb
anyway out of paranoia.

You need to identify every RO clone believed to be on the old server
and "vos remsite" that clone.

In general, I recommend simply avoiding this situation.  It's time consumin=
g.

While I can see why "synvldb" shouldn't put through a change if it
can't talk to an unreachable server (what if it's just a network
outage?) I do wish there was a "force" switch  to syncvldb to tell it
"no, really, the server you think the volume is on is really really
dead."

On the bright side, "changeloc" is the 2nd greatest thing since
toasted sliced bread.