You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by Christian Posta <ch...@gmail.com> on 2013/05/02 06:37:46 UTC
Re: Loads of TIME_WAIT connections
Check out http://www.unix.com/solaris/12257-time-wait-interval.html
On Tue, Apr 30, 2013 at 12:05 AM, lacigas <la...@gmail.com> wrote:
> Sorry if this has been asked before, I have seen several posts, but none of
> them helped.
>
> I'm running AMQ 5.8 on linux 64bit. Camel routes generated by Talend ESB.
>
> As a test I have one pooled connection factory (default values) and one jms
> consumer.
> The jms consumer is configured with cacheLevelName=CACHE_CONSUMER.
>
> Alone with this simple setup my camel route produces ~60 TIME_WAIT
> connections.
> Our system consists of several routes with multiple JMS components and if
> they are all running we have between 2000 and 3000 TIME_WAIT connections
> and
> that seems to kill our server.
>
> What I don't know if this behaviour is normal (loads of TIME_WAIT
> connections) and if not how to solve it.
>
> Thanks for your help.
>
> Laci
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/Loads-of-TIME-WAIT-connections-tp4666433.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>
--
*Christian Posta*
http://www.christianposta.com/blog
twitter: @christianposta
Re: Loads of TIME_WAIT connections
Posted by lacigas <la...@gmail.com>.
You are right, the problem is caused by the client.
The expiry timeout for the PooledConnectionFactory was set to 1ms by
default,
so every connection was going to TIME_WAIT state after 1ms if not used.
I have increased this numer and now I have a nice low number of TIME_WAIT
connections.
thanks
Laci
--
View this message in context: http://activemq.2283324.n4.nabble.com/Loads-of-TIME-WAIT-connections-tp4666433p4666566.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Loads of TIME_WAIT connections
Posted by Geurt Schimmel <GS...@schubergphilis.com>.
But the real problem is that a client in a tcp connection doesn't properly (usually not at all) close down the connection.
If you don't have control over clients (eg in a public network), lowering the time_wait interval is the only solution, but otherwise, look for missing close() calls at the client side.
Geurt
On 2 mei 2013, at 06:38, "Christian Posta" <ch...@gmail.com> wrote:
> Check out http://www.unix.com/solaris/12257-time-wait-interval.html
>
>
>
> On Tue, Apr 30, 2013 at 12:05 AM, lacigas <la...@gmail.com> wrote:
>
>> Sorry if this has been asked before, I have seen several posts, but none of
>> them helped.
>>
>> I'm running AMQ 5.8 on linux 64bit. Camel routes generated by Talend ESB.
>>
>> As a test I have one pooled connection factory (default values) and one jms
>> consumer.
>> The jms consumer is configured with cacheLevelName=CACHE_CONSUMER.
>>
>> Alone with this simple setup my camel route produces ~60 TIME_WAIT
>> connections.
>> Our system consists of several routes with multiple JMS components and if
>> they are all running we have between 2000 and 3000 TIME_WAIT connections
>> and
>> that seems to kill our server.
>>
>> What I don't know if this behaviour is normal (loads of TIME_WAIT
>> connections) and if not how to solve it.
>>
>> Thanks for your help.
>>
>> Laci
>>
>>
>>
>> --
>> View this message in context:
>> http://activemq.2283324.n4.nabble.com/Loads-of-TIME-WAIT-connections-tp4666433.html
>> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>>
>
>
>
> --
> *Christian Posta*
> http://www.christianposta.com/blog
> twitter: @christianposta