It's all about the answers!

Ask a question

core.dmp files hogging space on C:\


Matthew Clark (10711244) | asked Mar 29 '16, 2:27 p.m.
The C:/ on my CLM server has been filling up lately. I realized a handful of files within the WebSphere directory were responsible for about 63 GB of data.

The files are labeled "core.********.******.****.****.dmp"

The ****** portion seems to be a date/time stamp for when the file was created.
The files are stored in the following directory...

C:\IBM\Websphere\AppServer\profiles\AppSrv01.

Can anyone tell me about these files? They average about 12-14 GB in size.
Thanks!

3 answers



permanent link
Zeeshan Choudhry (6541612) | answered Mar 29 '16, 2:56 p.m.
these are system dmp files which are triggered on an event of running Native or Heap out of memory , depending on what you have set as -Xdump JVM argument

These files in huge in size , depending on how much heap is configured and how much was used when the process Ran out of memory,

I would expect your server not to work if that happens.

Otherwise check Xdump argument to see if has argument to write system dump when ever you take a JAVA core.


Comments
Matthew Clark commented Mar 29 '16, 3:06 p.m.

These are the Generic JVM arugments I have set in Websphere Admin Console....

-Xmx12g -Xms12g -Xmn3g -Xgcpolicy:gencon -Xcompressedrefs -Xgc:preferredHeapBase=0x100000000 -XX:MaxDirectMemorySize=1G -Dlqe.config.location=file:///C:\IBM\JazzTeamServer6.0.1\server\conf\lqe

I don't see anything about "-Xdump". Can you tell me where to look for that argument?  Thanks!


Donald Nong commented Mar 29 '16, 7:12 p.m.

If you're using WAS 8.5.5, it by default creates a system dump when the first OOM occurs.
http://www.ibm.com/support/knowledgecenter/SSAW57_8.5.5/com.ibm.websphere.nd.doc/ae/ctrb_java626.html

Before you decide to remove the dump files or change JVM settings, have a look at the first 10 lines of the javacore* files, and you should know why these dump files were created. OutOfMemory in the Java "heap space" can sometimes be easily resolved by bumping up the heap size, but other errors can be complicated and require further investigation to get a proper solution.


permanent link
Rafael Rodriguez Montes (23013126247) | answered Mar 29 '16, 2:59 p.m.
you can delete them, its safe to do it

permanent link
Alan Sampson (93749) | answered Mar 29 '16, 3:13 p.m.
JAZZ DEVELOPER
edited Mar 29 '16, 3:14 p.m.
Depending on their age (as you say, the file names contain timestamps), if they're very old then you can probably just delete them as too much time has likely passed for them to be of any value.  If however they're recent, it's an indication that your WebSphere server is running into problems and is crashing so you would do well to move them somewhere safe and begin an investigation into the issue that's causing the crashes.

They are often accompanied by javacores, heapdumps, snapdumps etc. All these files can be used to diagnose problems within the WebSphere server so if you do move them you should keep them together.

Comments
Matthew Clark commented Mar 29 '16, 3:23 p.m. | edited Mar 29 '16, 4:03 p.m.

Some of them are older (from 2015). Some of them are recent, including today, 2 days ago, and 5 days ago. There are also javacores and heapdump files that have similar timestamps so I will try to keep those files together (the recent ones anyway).

Is this a setting I have on somewhere in Websphere?  Can  I turn off the core dumps, javacores, and heapdumps?

It seems the timing of these dumps has to do with reboots of my CLM server. Either during shutdown, or bootup, something is crashing and causing these dump files.




Alan Sampson commented Mar 29 '16, 3:56 p.m. | edited Mar 29 '16, 3:58 p.m.
JAZZ DEVELOPER

I think the -Xdump option can do that but it's probably not a good idea to prevent these files from being created; without them you won't be able to determine the cause of the crashes and if you ever decide to open a PMR the WebSphere support team will need them to begin diagnosing the issue. As mentioned in another answer; the likelihood is your server is running out of RAM so you'd be better served in determining why the dumps are being taken.

I understand that the location that these files are written to can be controlled by environment variables so if space on the C:\ drive is at a premium you could consider redirecting them to a different drive instead.

There are a number of resources to help troubleshoot WebSphere crashes on-line from IBM and elsewhere. These links from IBM and StackOverflow look particularly useful:

http://www-01.ibm.com/support/docview.wss?uid=swg21181068

http://www-01.ibm.com/support/docview.wss?uid=swg21393547

http://stackoverflow.com/questions/7361233/how-to-analyse-websphere-core-dmp-file-and-snap-trc-files

http://stackoverflow.com/questions/19152376/how-to-stop-creating-heap-files-java-core-files-which-are-creating-in-jvm-insta

Your answer


Register or to post your answer.


Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.