[ic] interchange hanging

Dan Bergan dan at berganconsulting.com
Tue Jun 8 03:03:40 UTC 2010

On Sat, Jun 5, 2010 at 9:31 PM, Peter <peter at pajamian.dhs.org> wrote:
> On 06/06/10 12:23, Dan Bergan wrote:
>> If anyone could help me tracing the source of this problem, I would
>> really appreciate it:
>> Interchange 5.4.5
>> Periodically, my catalogs will become unresponsive.  When I ssh in to
>> my server, I see one interchange process taking all the cpu.  I don't
>> see anything out of the ordinary in the interchange logs, but in my
>> apache logs I see "premature end of script headers" for one catalog,
>> in another catalog the apache log says "ALERT: Interchange server not
>> running".
>> Restarting interchange will sometimes clear up the problem, but it
>> usually reappears a few times over an hour or so.  I'm guessing that
>> it is a bad robot or something similar, but I'm just not sure how to
>> track down what is causing this or what my solution should be.
> It is likely that you have an endless loop or an extremely long loop in
> one of your catalogs somewhere.  Track it down to the catalog with the
> problem first by eliminating one catalog at a time until the problem
> goes away, if you can't do this then set up a separate Interchange
> daemon process and move one catalog at a time over to it and you should
> be able to tell then.
> Once you've tracked down the catalog you'll want to put logging in that
> will (hopefully) tell you what page the problem is on.  Something like
> this *may* work in your catalog.cfg:
> Preload <<EOA
> [log]Starting: [data session last_url][/log]
> AutoEnd <<EOA
> [log]Ending: [data session last_url][/log]
> Wait for things to lock up and hopefully you'll get an entry in your
> catalog's error.log file with a "Starting:" line but no corresponding
> "Ending:" line.  Then you'll know the page and you can then track things
> down further from there.
>> I am planning on upgrading my hardware soon -- is there any way for me
>> to tell if it is a hardware bottleneck?
> Doubtful that it's a hardware bottleneck.
> Peter

Thanks Sam and Peter, I'll give those ideas a try.


More information about the interchange-users mailing list