[ic] Re: ALERT: bad pipe signal received for /page.html
Ron Phipps
ron at endpoint.com
Tue Dec 12 13:03:50 EST 2006
Music wrote:
>
> ----- Original Message ----- From: "Kevin Walsh" <kevin at cursor.biz>
> To: <interchange-users at icdevgroup.org>
> Sent: Tuesday, December 12, 2006 4:12 PM
> Subject: Re: [ic] Re: ALERT: bad pipe signal received for /page.html
>
>
>> "Music" <music at labyrinth.net.au> wrote:
>>> ifdef TRAFFIC =~ /rpc/i
>>> Message RPC traffic settings.
>>> PreFork Yes
>>> StartServers 5
>>> MaxServers 0
>>> MaxRequestsPerChild 100
>>> HouseKeeping 2
>>> PIDcheck 120
>>> ChildLife 30 minutes
>>> endif
>>>
>>> I also tried disabling ChildLife and PreFork and increasing start
>>> servers.
>>>
>>> All with no joy.
>>>
>>> 'Applying Changes' to one catalogue puts the Interchange daemon into
>>> faulty
>>> state.
>>>
>>> No errors are displayed in IC error.log, catalog error.log nor apache
>>> error
>>> logs.
>>>
>>> Just thought I would document this on the list to see if any feedback
>>> etc.
>>>
>> Just as an experiment, could you reduce the ChildLife to 60 (1 minute),
>> restart Interchange and reproduce the problem. I suspect that the
>> problem
>> will clear itself within one minute.
>
> Yes, spot on. The problem pages disappeared after the 60 seconds!
>
>>
>> If that happens then we know that the problem is probably in Interchange
>> itself, rather than than in Interchange::Link.
>
> Not using Interchange:Link on this server - just regular tlink CGI
> running in Inet Mode.
>
>>
>> The thinking is that the "apply changes" (or whatever), causes one of
>> the Interchange children to throw its toys out of its pram. The constant
>> refreshes then causes the naughty child to receive a request every now
>> and again, resulting in a moody page.
>>
>> One thing that may debunk my theory is your assertion that the problem
>> still occurs when PreFork is switched off. Can you check that again
>> for me.
>
> Testing with Pre-Fork - Oh dear - it appears Pre-For is the problem.
> I commented out Pre-Fork and ChildLife and the server is running very
> nicely.
> I can't reporduce the problem. Now I have to try and remember why I
> added Pre-Fork and ChildLife; I hope it is not too dramatic a jolt!
>
> I have dropped the PERL_SIGNALS=unsafe in restart command now also and
> all is running well.
You may have been running Pre-Fork mode to help with issues where sendmail or your gateway charges were not working properly. If that is the case make sure that MaxServers 0 is present in your high and low traffic areas of interchange.cfg, this is the new default.
--
Ron Phipps
End Point Corporation
ron at endpoint.com
More information about the interchange-users
mailing list