Re: open sleeps

[Brian May <brian@vpac.org>]

Dave Chinner wrote:
> This is just the stupid (broken) ia32 stack tracing - it outputs every
> single object it sees in the stack that may be function call. Hence
> you see some stuff that is from past stack traversals. I'd say this
> is sleeping in a select() call, not doing anything related to
> xfs or networking....
>   
Hello,

Thanks for your help.

Just a followup: Every indication seems to be that it is the firewall in 
the virus scanner we use, F-Secure, that prevents Windows from releasing 
oplocks properly. After turning off the firewall, the oplocks get 
released and there is no problem. Possibly related to the timing of wpkg 
being invoked before system logons too.

Might be time to consider another virus scanner/firewall solution.

Brian May