Rebooting the servers like this can cause split-brains. Example: there is a volume with two bricks brick-A on server-A, and brick-B on server-B. Lets say server-A is brought down and while it is down file-1 is modified which needs to be healed to server-A from server-B as soon as it is up. Now as soon as server-A is up lets say server-B is brought down before the healing on file-1 could complete. Now lets say file-1 is modified again. Now when server-B comes backup, file-1 would be in split-brain. So make sure the files are healed after a server is brought up then only reboot the other machine. Pranith ----- Original Message ----- > From: "Nux!" <nux at li.nux.ro> > To: "Pranith Kumar Karampuri" <pkarampu at redhat.com> > Cc: "Gluster Users" <gluster-users at gluster.org> > Sent: Monday, August 5, 2013 5:12:27 PM > Subject: Re: No active sinks for performing self-heal on file > > On 05.08.2013 12:35, Pranith Kumar Karampuri wrote: > > Thanks for the logs Nux. Now that the issue is fixed, let us try to > > figure out what lead to this situation for those two files. > > Could you let me know when the server is rebooted. > > Hi, well, this has been going on for a while and I don't remember when > it was done exactly, perhaps 1.5-2 weeks ago or so. Also worth noting > that after the first server was rebooted, as soon as it came online we > rebooted the other one. > This is about the only peculiar detail that I can think of right now. > Would this explain some of the problems that are going on? > > Lucian > > -- > Sent from the Delta quadrant using Borg technology! > > Nux! > www.nux.ro >
