[OpenAFS] Re: Afs User volume servers in VM's
Renata Maria Dart
renata@slac.stanford.edu
Wed, 26 Oct 2011 10:55:05 -0700 (PDT)
Hi Jeff, this is our old problem which we have suffered from
since day 1. It is multiple clients all writing to one file
or one directory at the same time. The threads on the server
plummet and waiting processes start backing up.
Renata
On Wed, 26 Oct 2011, Jeffrey Altman wrote:
>The issue is not going to be callbacks on a volume since callbacks are
>not issued on a per volume basis for read/write volumes and I assume
>that these jobs are not issuing volume releases which would trigger
>callback breaks on readonly volumes.
>
>The more likely scenario are:
>
>. file lock requests/releases on a specific file
>
>. data stores to a particular file
>
>. directory content changes (create, rename, link, unlink) in a single
>directory
>
>Jeffrey Altman
>
>
>
>