Dear all,
We have several sshd installations based on Cygwin, running on Server
2003 SP1 (and also SP2).
Over the last 12 months or so, there seems to have been a lot of
problems with sshd - the eventlogs cite a fork: resource temporarily
unavailable - error.
I've checked the obvious (ie port 22/tcp isn't already bound
elsewhere, ran filemon/regmon) but failed to find the reason. Is
there a way of logging _which_ resource is temporarily (or rather,
persistently) unavailable?
There seems to be a higher occurrence of this issue on Citrix PS
4.0/4.5 presentation servers, but after ruling out various memory and
cpu optimisations (eg DLL rebasing - I've tried the rebaseall script
and confirmed that none of the sshd or cygrunsrv dependencies have
been touched by Citrix MPS), I can't pin it down to a specific patch
level or configuration.
I've seen various other threads relating to this issue, but no
solution. I'd like to find the solution, but I seem to be missing
some vital pieces of the puzzle.
I have a nasty feeling this is going to involve learning a bit more
about understanding the drwtsn32.exe crashdumps (manually attaching
the debugger, as crashdumps aren't a symptom of this issue).
Thanks in advance - for any pointers/hints,