[ic] RPC mode not so reliable?
emailgrant at gmail.com
Sat Oct 27 17:39:20 EDT 2007
> > I've switched to RPC mode to prevent intermittent 500 errors in my
> > catalog, but now when I browse my catalog there comes a time when the
> > server will not respond. Restarting IC always fixes it. I'm on 5.2.
> > Is an upgrade the preferred solution, or is that unlikely to make a
> > difference? I'm using the following settings:
> > PreFork Yes
> > StartServers 5
> > MaxServers 0
> > MaxRequestsPerChild 100
> > HouseKeeping 2
> > PIDcheck 120
> > - Grant
> > _______________________________________________
> > interchange-users mailing list
> > interchange-users at icdevgroup.org
> > http://www.icdevgroup.org/mailman/listinfo/interchange-users
> Hello Grant,
> I can't remember if I've offered this advice to you before or not, but have you tried running in high traffic mode, but setting MaxServers to 0 (this is the default in the latest ICs)? Here are the settings we run for FrozenCPU:
> Variable TRAFFIC high
> ifdef TRAFFIC =~ /high/i
> Message High traffic settings.
> HouseKeeping 2
> #MaxServers 10
> MaxServers 0
> PIDcheck 600
> Do the IC or Apache logs saying anything about what the cause of those intermittent 500 errors might be? We've found that most versions of IC, running on the latest perl, to be stable.
I have tried high mode with MaxServers 0 but I still get the 500s. No
clues in error.log or error_log besides:
Connection reset by peer at /usr/lib/perl5/site_perl/5.8.8/Interchange/Link.pm
I do use Interchange::Link so maybe that has something to do with it.
Running in RPC mode did seem to fix it. I'd like to give it more time
to be sure, but IC is hanging and requiring a restart every 8 hours or
so under RPC mode. I use Gentoo and all of my software is completely
up to date except for IC.
More information about the interchange-users