Crucible 3.2.3 OOM issues

CTO Build, Run, and Architecture

Last month, we upgraded to Crucible 3.2.3, and started noticing a memory leak issue with the tool. The leak seems to be associated with Fisheye scanning a large repository. The leak causes the server to restart - we have the default panic_on_oom turned on. The large repo in question did not index to completion with previous versions of crucible. The indexing would stop midway or error out without completion.

Have you noticed anything similar? Any tips on how to avoid the issue.

gustavo_refosco

Hi,

It is not possible to determine exactly how much memory Crucible needs, as this would depend of things like size and complexity of repositories and its indexing operations, for example, so large and complex repositories will require sometimes a large amount of memory.

As you said this repository never got indexed, it seems it needs indeed more memory for it. The following document may be helpful in your case, as it points how to increase the memory available:

https://confluence.atlassian.com/display/FISHKB/Fix+Out+of+Memory+errors+by+increasing+available+memory

Attend to increase the -Xmx value, which is the maximum memory allowed for your Crucible instance.

You may also find valuable information in the following documents regarding the configuration of FishEye in order to get a better performance while indexing:

https://confluence.atlassian.com/display/FISHEYE/Best+practices+for+FishEye+configuration

https://confluence.atlassian.com/display/FISHEYE/Improve+FishEye+scan+performance