Hi Markus, Did you do anything to get that problem? i.e bring a server down , create a new file etc? I am trying to see what steps caused this bug. Krishna On Fri, Jul 18, 2008 at 2:42 PM, Markus Gerstner <m.gerstner at bmiag.de> wrote: > Heya Mailing-List, > > we're using an AFR'd Namespace with one subvolume on each server in the FS, client-configuration as follows: > volume backup-afr-ns > type cluster/afr > subvolumes backup-ns1 backup-ns2 backup-ns3 backup-ns4 > end-volume > > The namespace-volumes are stored on XFS, clients and servers use debian. XFS is mounted with "noatime"-Option. > Data-Volumes are stored on ext3 with noatime and user_xattr. > > We set it up like that to ensure redundancy on the namespace without being dependant on the clients. Maybe this is not the proper way to use a namespace, if so, please let me know. > > Should it be the proper way, I'd like to ask why the date on one directory suddenly changed to 1970-01-01. The date is still correct in the data-volumes of the servers but set to 1970 on all servers' namespace-volumes. Unfortunately, the "fix" from the devel-mailinglist, using "head -1 * > /dev/null" does not apply here for obvious reasons. > > Is there a way to resync the namespace and reset the date or - even better - a way to prevent this from happening again? :-) > > Regards, > Markus > > > _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users >
