[
https://issues.apache.org/jira/browse/AXIS2C-858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12555208
]
Manjula Peiris commented on AXIS2C-858:
---------------------------------------
I did some testing with the mtom sample in Axis2/c and Apache 2.6 . I consumed
the mtom service in a loop with 3 shells . First I tested it with 4 MB
attachments and noticed a significant memory growth. Following are some
observations in my machine.
1. Before the loop memory usage is 29%, and it increased to 70% immediately
after starting the loop. And it will grow up to 90% memory usage and request
handling became very slow. The other thing is when I stopped the loop the
memory usage is remained same and does not decreased. It only decreased when I
stopped Apache server.
2. I did the same with simple axis2 server and memory usage remained between
40-43%. And there is no significant memory growth at the start of the loop.
3. I also tested Apache2 with a small attachment (<< 4MB) and the memory usage
increased gradually , But not a significant increase at the start of the loop.
BTW I didn't notice any crash in the server. But i think there is some memory
leak in the axis2 httpd module. Valgrind output does not show any thing but I
will try to find where the problem is.
> Apache runs out of memory and crashes with a simple mtom service test run in
> a loop
> -----------------------------------------------------------------------------------
>
> Key: AXIS2C-858
> URL: https://issues.apache.org/jira/browse/AXIS2C-858
> Project: Axis2-C
> Issue Type: Bug
> Components: core/engine
> Affects Versions: 1.1.0
> Environment: Linux 2.6 kernel
> Reporter: Subra A Narayanan
> Assignee: Manjula Peiris
>
> I have a simple mtom service written in Axis2/C which receives requests from
> clients with attachments. I wrote a client to consume this service and ran a
> test in a loop with 4 mb attachments. I noticed that the apache server ran
> out of memory and became very unresponsive after about 45 mins and then core
> dumped. The server response seemed to improve after some of the
> threads/processes died due to lack of memory and new threads were created to
> process the request.
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
