Retrigger failed analysis?


(Björn Kautler) #1

Is there a way to retrigger a result analysis?
The background job failed with java.lang.OutOfMemoryError: Java heap space
So I’d like to increase ce heap space and retrigger the analysis without the need to run the scanners again.


(Björn Kautler) #2

Actually the analysis failed, but some information made it in anyway, I see changed numbers compared to before the analysis. While the last analysis date in the upper right is still the one from before the analysis. This also doesn’t sound quite right.


(G Ann Campbell) #3

Hi,

Unfortunately, there’s no way to recover this failed background task. You’ll have to (allocate more memory, bounce the instance, and) reanalyze. The inconsistent state is known. I won’t call it a “known issue” because that’s just the way background task processing is structured: we do several commits throughout, so a background task that fails part way through does leave the project in an inconsistent state until the next analysis. I asked some questions about this this morning and it seems that it’s not obvious to move to a single commit at the end because, as you saw, we’re sometimes dealing with some very large data sets. Anyway, this quetion is on the mental radar (no tickets at this point) and may be addressed in the future.

 
HTH,
Ann


(Björn Kautler) #4

Ok, thanks for the info.
I had a DB dump of right before the analysis, so I played it back in, increased the CE memory from 1G to 2G and redid the analysis.
This time it went through.