All Status is "S" Tomcat manager

[Jian Wu <fa...@gmail.com>]

Hi,

I am running a website with a load balancer (software) and three web
servers, each of which deploys an instance of Tomcat 7. While the other two
web servers look OK, there is one web server which often gives me trouble
after restarting Tomcat.

Basically, the maxThread (2000) is reached and the server is almost not
accessible. When I open the server:8080/manager of that server, I found
that all the Stages under "http-nio-8080" are "S"! The values of "BSent"
and "BRecv" are all 0 KB. What does that mean? To solve this problem, the
only way I know is to restart Tomcat, but after several days, the problem
emerges again.

These three servers are deployed in exactly the same way, so it is very
strange that one of them uses all 2000 theads, but the others use only less
than 300 threads! Currently, the load balancer distributes 3/7 traffic to
Server 1, 2/7 traffic to Server 2, and 2/7 traffic to Server 3.  The
problematic server is Server 2.

Thanks
Jian Wu

Re: All Status is "S" Tomcat manager

[Christopher Schultz <ch...@christopherschultz.net>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Jian,

On 4/24/15 10:17 AM, Jian Wu wrote:
> I am running a website with a load balancer (software) and three
> web servers, each of which deploys an instance of Tomcat 7.

Just to be clear, you have a single load-balancer, 3 web servers
(httpd?), and 3 Tomcat servers? Or is Tomcat considered a "web server"
in this scenario?

> While the other two web servers look OK, there is one web server 
> which often gives me trouble after restarting Tomcat.
> 
> Basically, the maxThread (2000) is reached and the server is almost
> not accessible.

That's a lot of threads.

> When I open the server:8080/manager of that server, I found that
> all the Stages under "http-nio-8080" are "S"! The values of
> "BSent" and "BRecv" are all 0 KB. What does that mean?

Those are metrics I don't recognize. Are these stats kept and shown by
the load balancer?

> To solve this problem, the only way I know is to restart Tomcat,
> but after several days, the problem emerges again.
> 
> These three servers are deployed in exactly the same way, so it is
> very strange that one of them uses all 2000 theads, but the others
> use only less than 300 threads! Currently, the load balancer
> distributes 3/7 traffic to Server 1, 2/7 traffic to Server 2, and
> 2/7 traffic to Server 3.  The problematic server is Server 2.

Can you post your configuration for Tomcat's <Connector>s? Remember to
remove any sensitive information that may be in there.

- -chris
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
Comment: GPGTools - http://gpgtools.org
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=eRCX
-----END PGP SIGNATURE-----

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org