I am at the moment just testing AFS, but this would be the eventual
production configuration, the idea was to have multiple geographically
disparate openafs servers to act as redundant filesystem storage in the
event of any one system failure, do you think it would actually be less
stable to use AFS than manually rdist'ing the production files around
the server ring?
Regards
Eric
Chris Huebsch wrote:
On Sat, 7 May 2005, Eric Bennett wrote:
You're right, /vicepa on corvus went boom, I have no idea how or why,
but it's definitely gone, it was a loopback file from a local
filesystem and when the system was rebooted the original file
disappeared entirely.
A Loopback-file? I hope, you are just testing AFS...
If you want to "fake" a partition you can use an empty file named
"AlwaysAttach" in the /vicepX-directory. It will be used as a partition
even if its not a true partition.
Is it possible to copy the volumes from raven or do they need to be
remade in their entirety? I've tried vos copy root.afs raven /vicepa
root.afs corvus /vicepa -verbose but it just says the volume already
exists, so I then tried a syncvldb thinking that perhaps the vldb
needed to know there was a problem but it reported successful and
still lists corvus' old volumes in listvldb.
Anyone have any ideas?
There are a lot of commands, which manipulate the vl-db (just do a vos
help).
My favourite command would be "vos convertrototw" (or like that).
If your version does not have this command, try "vos dump" and "vos
restore" to dump your volumes into a file and restore them as
rw-volumes.
Good luck
Chris
smime.p7s
Description: S/MIME Cryptographic Signature
