On our most heavilly loaded Samba server (running 1.9.18p10 under Solaris
2.5.1) we have got occaisional meltdowns with load average rocketing to 100+.
The evidence indicates lots of smbd processes running. Syslogs indicate lots
of:
Nov 12 12:48:35 zeus.brunel.ac.uk smbd[21572]: Failed to set socket option
SO_KE
EPALIVE
Nov 12 12:48:35 zeus.brunel.ac.uk smbd[21572]: Failed to set socket option
TCP_N
ODELAY
Nov 12 12:48:41 zeus.brunel.ac.uk smbd[21573]: Failed to set socket option
SO_KE
EPALIVE
Nov 12 12:48:41 zeus.brunel.ac.uk smbd[21573]: Failed to set socket option
TCP_N
ODELAY
Also on the most recent incidence (happening once or twice a week), it was
reportedhed that there were a lot of smbd connections bound to socket *.139 in
the idle state.
A glance at the source code indicates such failures ignored.
Has anyone else experienced this?
What more should I do to diagnose?
Thanks,
--
-----------------------------------------------------------------------------
| Peter Polkinghorne, Computer Centre, Brunel University, Uxbridge, UB8 3PH,|
| Peter.Polkinghorne@brunel.ac.uk +44 1895 274000 x2561 UK |
-----------------------------------------------------------------------------
