Lindsay Mathieson
2016-Apr-20 06:22 UTC
[Gluster-users] 3.7.11 - Brick died, can't restart
On 20 April 2016 at 16:15, Bishoy Mikhael <b.s.mikhael at gmail.com> wrote:> try restarting glusterd. > # service glistered restart > > if it didn?t work, try killing glusterfsd PID(s) > # kill $(ps -ef | grep glusterfsd | awk '{print $2}?) > t > hen, restart glusterd > # service glusterd restartAlready tried all those, its a zombie linux process with parent pid 1, so can't be killed short of a reboot. It seems to have released the socket handle now (49156) but the brick still isn't connecting to the others. Going to have to reboot. -- Lindsay
Lindsay Mathieson
2016-Apr-20 07:55 UTC
[Gluster-users] 3.7.11 - Brick died, can't restart
On 20 April 2016 at 16:22, Lindsay Mathieson <lindsay.mathieson at gmail.com> wrote:> Already tried all those, its a zombie linux process with parent pid 1, > so can't be killed short of a reboot. > > It seems to have released the socket handle now (49156) but the brick > still isn't connecting to the others. Going to have to reboot.Reboot sort of fixed the problem and revealed the probable cause - hard disk failed. Never listen to people saying that desktop drives can be used in server devices ... this is the 2nd WD Black thats failed less than a year after installation. And the trouble with them is they don't support TLER,, so when they fail they bring the whole server down. Pre-emptively replacing them all with WD Reds. Data is safe though - RAID!) mirror and Gluster Rep 3 :) VM's on the other nodes didn't even glitch. -- Lindsay