Displaying 2 results from an estimated 2 matches for "filur05".
Did you mean:
filur00
2018 Dec 03
3
Samba 4.9.3 and the "10 hour problem"
...the trouble to occur at 2pm - not good…)
This production server had around 400 concurrent users when at the time it failed this time.
In order to get it to start servicing users again I tried to kill the “Winbindd” daemons using “pkill winbindd” and all but one of the daemons died:
> root at filur05# ps auxww | egrep winbind
> root 1550 0.0 0.0 271956 127236 - S 07:12 0:23.66 winbindd: domain child [AD] (winbindd)
> root 29659 0.0 0.0 237240 122108 - I 10:04 0:00.02 winbindd: domain child [FILUR05] (winbindd)
> root 99774 0.0 0.1 267...
2018 Dec 03
2
Samba 4.9.3 and the "10 hour problem"
Hmm…
I see that there is a patch in the bugzilla page for that bug. I guess I could try that one… :-)
I notice in the patch that there is a lot of talk about SMB2 - we use SMB3 mostly now. But perhaps there is some code sharing? Or perhaps SMB2 is used when talking to the AD servers? Or the few SMB2-talking clients causes the problem for all of the other users? When testing the “timeout” issue I