tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Samad - Yieldbroker <Alex.Sa...@yieldbroker.com>
Subject RE: IIS 7.0 Worker process crashes on App Pool recycling since ISAPI redirector 1.2.33
Date Sun, 18 Mar 2012 20:08:49 GMT
[snip]

> > Just wondering (thinking a bit left field), but  question was asked of me.
> Any reason to have a shm across processes.
> >
> 
> So that worker state is replicated across processes.
> Eg, if connection to one backend is broken each process will have to discover
> that by itself. Shared memory allows that if one worker process discovers
> that a backend is down other won't have to go trough the lengthy process of
> discovering that.
> 
> Also loadbalancer load factors are calculated across all processes not just for
> the current one. So, the benefits are real.

Not deny that, just thinking out aloud, also been asking myself if there is any reason  to
run a 4 process web garden. The only benefit I can see is that if 1 process dies there will
be other process to continue processing.

1 process + threads v's X processes + threads


> 
> However unlike with httpd, IIS does not offer parent/child concept so we
> don't have the shared memory 'controller'.
> The ultimate solution would be to start a separate controller process, but
> that's probably something for the future.
> 
> Even if the current set of patches work, I'm more convinced we should have
> a directive to disable shared memory usage, and think I'm going to add that
> option for 1.2.34 anyhow.

So would that effectively do what I was suggesting above ?


Thanks
> 
> 
> Regards
> --
> ^TM
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
> For additional commands, e-mail: users-help@tomcat.apache.org

Mime
View raw message