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

Louis,

On 9/26/18 14:42, Louis Zipes wrote:
> Hi all, Tomcat 7.0.54 running on Windows 2012
> 
> We are running a third party application on Tomcat and today we
> have intermittently run in issues where the application stops
> working.  The big changes in our system is that we have added more
> end users and we are at year end so of course everyone is hitting
> the system hard. Even if we force a log out of all users and stop
> all background jobs then the application doesn't recover.
> 
> We see no active sessions on the database (our application is
> connecting to an Oracle database) and I see no clear error messages
> in either our third party application logs or the Tomcat logs (ex.
> OutofMemory).  When we go to the Windows Task Manager we did not
> see the machine's Memory max'd out but admittedly I didn't look at
> the Java session to see if was reaching its Heap Max.  The only
> thing that we noticed was that TCP connections went down right
> after the restart.  I did open up Jconsole under Java and I did
> force a garbage collection but that didn't seem to help.
> 
> We do have an Oracle Grid Control and we did get an alert in
> regards to Metric: [HTTP Transaction] Perceived Time per Page going
> past thresholds but not sure if that was just an old alert with and
> old range that was set up a long time ago or is a really valid
> clue.    Since this is PRD we had to get it back up and running so
> all I did was increase the Tomcat Xmx Heap size and restarted.  I'm
> not really confident that is the solution since as mentioned you
> tend to see a clear out of memory error if it was too small.
> 
> So a few questions:
> 
> 
> 1)     Does this sound like a known issue with this earlier version
> of Tomcat?

No.

> 2)     Should I turn up any logging on Tomcat and if so which
> ones?

Not yet.

> 3)     We didn't do a JSTACK dump while it was happening.  Would
> that have been useful?

Absolutely.

> 4)     Do we need to play around with MaxThreads and/or
> MaxConnections.  We do have maxThreads in our server.mxl but in DEV
> when we turned it down to a value = 5  hoping to overwhelm it
> nothing bad happened.

Don't change anything, yet.

> Once again, we are limited to what we could do and collect since it
> was PRD and we needed to restart it.  We restarted the Tomcat
> service and everything is processing fine for right now.  I will
> note that that we did have that bad Windows patch that prevented it
> from stopping and starting cleanly
> (https://stackoverflow.com/questions/51498291/tomcat-lockup-on-shutdow
n)
> but we have taken the break fix patch and the daily restarts seem
> to be fine since then.
> 
> Any comments/suggestions are appreciated!

Your most likely problem is database connection pool mismanagement:
connections aren't properly released and the pool empties. All threads
are left waiting on available database connections which will never be
replenished.

- -chris
-----BEGIN PGP SIGNATURE-----
Comment: Using GnuPG with Thunderbird - https://www.enigmail.net/
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=Zavb
-----END PGP SIGNATURE-----

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

Reply via email to