You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by Michael Molloy <mm...@ncycles.com> on 2003/01/08 22:54:33 UTC
Tomcat oddities
My company is in the deployment stage of a project that uses tomcat to
serve information from an oracle database to about 25 people. When the
app goes live, there will be about 150 people connected at any one
time. Tomcat 4.1.12 is running behind Apache on Windows 2000 on a
single cpu box, and Oracle is running on a separate Windows 2000 2-way
box. (Windows was the client decision, not ours.)
Issue One
We're seeing two serious issues, the first of which happens about once
a day, sometimes more. When there are several users on the system,
maybe up to 15, there is a freeze during which no one can get any
responses back from Tomcat. This period usually lasts from 5 to 15
minutes, after which the system returns to normal and everything zips
along.
We've tuned queries, so we don't think that is the problem. There may
still be a rogue query out there causing problems, but we think it's
unlikely. Besides, I don't know why that would stop everyone, which is
what's happening.
Another possibility that we've discussed is that tomcat is simply using
all of the memory and everything basically stops until GC occurs. This
seems the most likely to me, but I wanted to ask the group. I don't
know what the page file size is, nor do we have access to the server so
we can't check task manager.
Can anyone think of any other possibilities?
Issue Two
We've had two reports where a user has had data from an old session
show up in his/her current session. For example, I wrote a class that
stores information from 14 different JSPs. The object is put into the
user's session. In these two occasions, the user entered a new record
using these screens & saved the data to the database. The user took an
hour lunch break, which would have been long enough to timeout (set at
20 minutes), returned, and queried a different record. Some of the data
from the previous record showed up in the holder class in his/her
current session and was saved to the database.
Any idea what that could be? The only thing I can think of is that IE
is doing some data caching & mixing things up a bit.
Any suggestions will be greatly appreciated. We really need to get this
figured out quickly.
Thanks
--Michael Molloy
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: One other freaky thing
Posted by Jason Koeninger <jk...@jjcc.com>.
I've seen the same thing on Tomcat 3.3.x. Our upgrade to Tomcat 4.0.6
seems to have corrected it.
Good luck...
Best Regards,
Jason Koeninger
J&J Computer Consulting
http://www.jjcc.com
On Wed, 8 Jan 2003 19:59:25 -0500 (EST), Jason Pyeron wrote:
>We had a similar problem running tomcat 3.x (3.3?) and apache 1.x (1.19?)
>
>mod_jk sometimes would deliver static html/gfx from one context to
>another. Also there would be session corruption.
>
>To the best of my knowledge there was a on-line casino which had the same
>problem.
>
>I was never able to track this down. My only solution was to connect
>directly to tomcat.
>
>-jason pyeron
>
>
>On Wed, 8 Jan 2003, Michael Molloy wrote:
>
>Our application is running on a server in Pennsylvania. A user there
>was working as well as a user in Tennessee. The user in Tennessee got
>an error on a page, hit her back key, and the user in Pennsylvania's
>screen showed up on the Tennessee user's screen. The people in
>Tennessee are connected to the Pennsylvania system via a frame relay.
>
>Everything is contained within each user's session, so this should
>never happen. The application has been under development for a year
>now, and this has never happened before.
>
>Some kind of weird bug that we shouldn't worry about, or something that
>someone else has encountered?
>
>Thanks for any help,
>--Michael Molloy
>
>
>--
>To unsubscribe, e-mail: <ma...@jakarta.apache.org>
>For additional commands, e-mail: <ma...@jakarta.apache.org>
>
>
>--
>-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
>- -
>- Jason Pyeron http://www.pyerotechnics.com -
>- Owner & Lead Pyerotechnics Development, Inc. -
>- +1 410 808 6646 (c) 500 West University Parkway #1S -
>- +1 410 467 2266 (f) Baltimore, Maryland 21210-3253 -
>- -
>-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
>
>This message is for the designated recipient only and may contain
>privileged, proprietary, or otherwise private information. If you
>have received it in error, purge the message from your system and
>notify the sender immediately. Any other use of the email by you
>is prohibited.
>
>
>
>
>--
>To unsubscribe, e-mail: <ma...@jakarta.apache.org>
>For additional commands, e-mail: <ma...@jakarta.apache.org>
>
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: One other freaky thing
Posted by Jason Pyeron <ja...@pyeron.com>.
We had a similar problem running tomcat 3.x (3.3?) and apache 1.x (1.19?)
mod_jk sometimes would deliver static html/gfx from one context to
another. Also there would be session corruption.
To the best of my knowledge there was a on-line casino which had the same
problem.
I was never able to track this down. My only solution was to connect
directly to tomcat.
-jason pyeron
On Wed, 8 Jan 2003, Michael Molloy wrote:
Our application is running on a server in Pennsylvania. A user there
was working as well as a user in Tennessee. The user in Tennessee got
an error on a page, hit her back key, and the user in Pennsylvania's
screen showed up on the Tennessee user's screen. The people in
Tennessee are connected to the Pennsylvania system via a frame relay.
Everything is contained within each user's session, so this should
never happen. The application has been under development for a year
now, and this has never happened before.
Some kind of weird bug that we shouldn't worry about, or something that
someone else has encountered?
Thanks for any help,
--Michael Molloy
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
--
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
- -
- Jason Pyeron http://www.pyerotechnics.com -
- Owner & Lead Pyerotechnics Development, Inc. -
- +1 410 808 6646 (c) 500 West University Parkway #1S -
- +1 410 467 2266 (f) Baltimore, Maryland 21210-3253 -
- -
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
This message is for the designated recipient only and may contain
privileged, proprietary, or otherwise private information. If you
have received it in error, purge the message from your system and
notify the sender immediately. Any other use of the email by you
is prohibited.
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
One other freaky thing
Posted by Michael Molloy <mm...@ncycles.com>.
Our application is running on a server in Pennsylvania. A user there
was working as well as a user in Tennessee. The user in Tennessee got
an error on a page, hit her back key, and the user in Pennsylvania's
screen showed up on the Tennessee user's screen. The people in
Tennessee are connected to the Pennsylvania system via a frame relay.
Everything is contained within each user's session, so this should
never happen. The application has been under development for a year
now, and this has never happened before.
Some kind of weird bug that we shouldn't worry about, or something that
someone else has encountered?
Thanks for any help,
--Michael Molloy
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
RE: Tomcat oddities
Posted by "Noel J. Bergman" <no...@devtech.com>.
> Tomcat 4.1.12 is running behind Apache on Windows 2000 on a
> single cpu box, and Oracle is running on a separate Windows
> 2000 2-way box. (Windows was the client decision, not ours.)
Have you checked the 4.1 change log to see if there are fixes past 4.1.12
that might impact you?
> When there are several users on the system, maybe up to 15,
> there is a freeze during which no one can get any responses
> back from Tomcat. This period usually lasts from 5 to 15
> minutes, after which the system returns to normal and everything
> zips along.
That doesn't sound like GC (not for 5 - 15 min). That sounds like a
resource lockup hanging things until it times out. Is it possible that
you're locking up on the database? Which JVM and JDBC driver?
> nor do we have access to the server so we can't check task manager.
Might I suggest that you fix that situation?
> We've had two reports where a user has had data from an old session
> show up in his/her current session.
Are you *sure* that the session was terminated? You might want to use the
TomcatTestListener class that I posted last year to watch what happens. The
sort of thing you are describing is similar to why I wrote it.
--- Noel
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>