On Sun, Jan 30, 2011 at 11:28 PM, G. Clifford Williams < [email protected]> wrote:
> > My apologies if this has been addressed already on the list, I don't get to > keep up as much as I'd like. > > I've got several clients that manage config files via fossil (they're > moving from git). The repos live on a network share (either NFS or CIFS). > The problem is that when the fossil is opened from a second machine (one > that did not initially create the repo), accessing the networked volume, it > throws the following error: 'repository does not exist or is in an > unreadable directory:'. We know that we can get around this error with 'rm > _FOSSIL_; fossil open <fossilfile>' but it'd be great to not have this pop > up on repos that are either copied (as in rsync, scp, cp, etc.. ) or > accessed from multiple machines via a network share. > Posix advisory locking tends to give problems on many NFS implementations. You can cause Fossil to use dot-file locking instead by setting an environment variable: export FOSSIL_VFS=unix-dotfile That might fix your problem. Caution, though: If you have two different copies of Fossil running and one is using dotfile locking and the other posix advisory locking, then they might step on one another and corrupt the database! So if you use FOSSIL_VFS, make sure you use it consistently! > > Thanks > > _______________________________________________ > fossil-users mailing list > [email protected] > http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users > -- D. Richard Hipp [email protected]
_______________________________________________ fossil-users mailing list [email protected] http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
