Re: [noc] ** PROBLEM alert - wiki.squid-cache.org/HTTP is CRITICAL **

On Mon, Mar 30, 2009 at 10:42 AM, Amos Jeffries <squid3_at_treenet.co.nz> wrote:
> Kinkie wrote:
>>
>> On Mon, Mar 30, 2009 at 1:31 AM, Nagios
>> <nagios_at_nagios.measurement-factory.com> wrote:
>>>
>>> ***** Nagios  *****
>>>
>>> Notification Type: PROBLEM
>>>
>>> Service: HTTP
>>> Host: wiki.squid-cache.org
>>> Address: wiki.squid-cache.org
>>> State: CRITICAL
>>>
>>> Date/Time: Sun Mar 29 16:31:55 PDT 2009
>>>
>>> Additional Info:
>>>
>>> Connection refused
>>
>> Sorry guys.
>> Root cause of this was eu getting a RAM upgrade (4 gigs now). Upon
>> reboot, the squid acting as a rproxy failed starting up with this
>> error:
>>
>> 2009/03/29 19:20:19| Configuring Parent squid-cache.org/80/0
>> 2009/03/29 19:20:19| WARNING: Disk space over limit: 275488 KB > 102400 KB
>> 2009/03/29 19:20:19| assertion failed: ../../lib/Profiler.c:213:
>> "cstack[cstack_head].timer == type"
>>
>>
>> Squid Cache: Version 3.1.0.6-BZR
>> configure options:  '--enable-debug-cbdata'
>> '--enable-xmalloc-statistics' '--disable-wccp' '--disable-wccpv2'
>> '--disable-htcp' '--enable-ssl' '--enable-stacktraces'
>> '--enable-cpu-profiling' '--enable-storeio=ufs,aufs'
>> '--enable-disk-io=AIO,Blocking,DiskThreads'
>> '--enable-removal-policies=heap' '--prefix=/usr/local/squid'
>> '--enable-ltdl-convenience' '--disable-optimizations'
>> '--with-default-user=squid' '--cache-file=/tmp/config.cache'
>> '--with-filedescriptors=1024' '--with-valgrind-debug'
>> --with-squid=/home/kinkie/squid/squid-3.1-bzr
>>
>>
>
> Try killing the cpu-profiling for now. It's known to have problems
> sometimes. This seems to be one.

There seems to be more than that.
3.1.0.6-r9477 could not parse a cache_dir aufs line (yes, aufs was
compiled in), nor interpret the http_access lines.

It's BADLY in need of some checking (read: it's BROKEN).
For now I've had to bypass squid entirely, we're now running on apache
alone until we figure out what's wrong.

-- 
    /kinkie