tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Darren Kukulka" <>
Subject RE: Tomcat 6 Clustering issue
Date Tue, 08 Sep 2009 12:40:19 GMT
The clustering and session failover appears to work OK.

It's just that when one node is failing to respond in a timely fashion
to requests, the balancer does not appear to redirect some (or in some
cases all) requests to the good node.

As the balancer is configured on 'availability', what does that actually
mean in this case?  Availability of Tomcat on a node, or the
availability of the service to respond to requests?

-----Original Message-----
From: Sumedh Sakdeo [] 
Sent: 08 September 2009 10:49
To: Tomcat Users List
Subject: Re: Tomcat 6 Clustering issue

Hi Darren,
         For our setup tomcat 6 is able to handle failover to healthy
with proper session replication. I assume you have loaded load balancer
module, updated, made cluster changes to server.xml of
tomcat 6. My problem is next step of yours.

         In terms of monitoring you can use jkstatus of apache server.
jkstatus Web page to monitor web server requests.For tomcat clustering


On Tue, Sep 8, 2009 at 2:49 PM, Darren Kukulka <> wrote:

> Hi All,
> I've come across some interesting scenarios with a clustered Tomcat6
> environment and I'm wondering what the most common (or best) approach
> configuration would be.
> Basically, in a 2 node cluster, with a front-end Apache Web Server
> mod_jk balancer (on availability, with equal weighting) - when one
> decides to stop servicing requests, it appears that the front end may
> still be trying to send requests to it and consequently timing out, or
> taking a very long time to resend the request to the healthy node.
> In situations like this, the failing node is not down.  Nor is it low
> resources or experiencing thread blocking.  There's no immediate
> evidence of memory leaks but this will require further investigation.
> It just stops servicing requests.  Indeed, even when we try and
> interrogate the JVM sometimes  with JConsole, we cannot connect to the
> environment.
> Anyway, the crux of the issue is how can we configure our environment
> such that the healthy node will service all requests if the bad node
> degraded, although it is up?  There is also the question as to how we
> monitor for such scenarios.
> Cheers,
> Darren
> Connaught plc is a FTSE 250 company. We are the UK's leading provider
> integrated services operating in the compliance, social housing and
> sector markets.
> Please visit our website to see a full list of Connaught's Registered
> Companies
> Disclaimer:
> The information transmitted is intended only for the person or entity
> which it is addressed and may contain confidential and/or privileged
> material. Any review, retransmission, dissemination or other use of,
> taking of any action in reliance upon, this information by persons or
> entities other than the intended recipient is prohibited. If you
> this in error, please contact the sender and delete this message.
> Connaught plc, Head Office 01392 444546

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message