Re: [squid-users] unlinked squid process

Hi

we are using SQUID 2.6.STABLE13

we usually restarting squid by flushing it
service squid restart
service squid flush

flush)
         $0 stop
         sleep 2
         echo -n 'Flushing squid cache: '
         echo "" > /var/spool/squid/cache/swap.state
         echo_success
         echo
         /usr/local/squid/sbin/squid -z
         $0 start
         ;;

Quoting Amos Jeffries <squid3@treenet.co.nz>:

> squid@unixplanet.biz wrote:
>> yes, that is the squid which has one process left over
>> when we restart it.
>>
>> is there any document related to that shut down bug ?
>>
>
> I was thinking of #1529 a 3.0-pre bug sorry not 2.6*. Sounds like a
> similar effect despite the version.
>
> Which squid are you using?
>
> Are you closing squid through a custom script sending SIGHUP/kill to a
> pid or a ports script provided by an official maintainer?
>
> A custom script may still encounter problems if it handles RunCache
> badly. Could be KILL'ing the child squid worker instead of HUP'ing the
> RunCache control process.
>
> Amos
>
>> Thanks
>>
>> Quoting Amos Jeffries <squid3@treenet.co.nz>:
>>
>>> squid@unixplanet.biz wrote:
>>>> Just wanted to know the meaning of MISS_TCP.
>>>> I thought that might be an error message.
>>>
>>> Errors, as in 'problems with squid', shop up in cache.log
>>> and start with "WARNING:" or "fatal:" maybe followed by a "Starting
>>> Squid" if it was really bad.
>>>
>>>> Thanks for the document.
>>>>
>>>> I am going to install a squid log analyzer to see what objects
>>>> are cached ( I think we are caching quarries which I don't think
>>>> it is a good idea )
>>>
>>> Caching quarries could be a problem with all the rocks.
>>>
>>> Caching queries however, is how the web works faster than the network
>>> underneath it.
>>>
>>>> which one do you recommend ?
>>>> there is plenty of them but I was thinking of squeezer.
>>>>
>>>> is there a sequence in restarting apache and squid while
>>>> apache is configured to be the parent ?
>>>
>>> no.
>>>
>>>> because some time when we restart squid or apache the squid
>>>> process is not killed, and new process will be started.
>>>
>>> do you mean apache is started twce?
>>> or do you mean squid is started twice?
>>>
>>> the second maybe related to a shutdown bug, long fixed in 2.6s12+.
>>>
>>> Amos
>>>
>>>
>>>>
>>>> Quoting Colin Campbell <sgcccdc@citec.qld.gov.au>:
>>>>
>>>>> Hi,
>>>>>
>>>>> On Tue, 2007-10-23 at 01:29 -0400, squid@unixplanet.biz wrote:
>>>>>> when I do "ps -ef | grep squid" I see
>>>>>> squid 25921 25919 0 21:04 ? 00:00:00 (unlinkd)
>>>>>>
>>>>>> why it is unlinked ? I couldn't find any document to explain.
>>>>>
>>>>> It isn't "unlinkEd", it's "unlinkd" which is a daemon used to remove old
>>>>> objects from the cache.
>>>>>
>>>>>>
>>>>>> also I am using squid for accelerating my web server.
>>>>>> I see following message in my access log files.
>>>>>> TCP_MISS:FIRST_UP_PARENT "-"
>>>>>>
>>>>>> would you please let me know waht that message means
>>>>>
>>>>>> From the FAQ (http://wiki.squid-cache.org/SquidFaq/SquidLogs)
>>>>>
>>>>> TCP_MISS = The requested object was not in the cache.
>>>>> FIRST_UP_PARENT = The object was fetched from the first parent in the
>>>>> list of parents.
>>>>>
>>>>>> and if there is any thing I should concern about
>>>>>> TCP_MISS:FIRST_UP_PARENT "-" 0 !!!!
>>>>>
>>>>> It's hard to tell exactly what you're concerned about. Show more of the
>>>>> log entry you're worried about.
>>>>>
>>>>> Colin
>>>>> --
>>>>> Colin Campbell
>>>>> Unix Support/Postmaster/Hostmaster
>>>>> Citec
>>>>> +61 7 3227 6334
>>>>>
>>>>
>>>>
>>>>
>>
>>
>>
Received on Tue Oct 23 2007 - 08:19:35 MDT