OpenMS: FeatureFinderMetabo returns code 137

Hi everyone!

So, I am trying to process 4 files with FeatureFinderMetabo in combination with Parallel Chunk Start and Parallel Chunk End. The 4 files have a size of ~600Mb two of them and ~200Mb the other two. The workflow fails and I get the following error for 1 of the 4 branches of the parallel execution while the other three finish successfully:

ERROR FeatureFinderMetabo 0:673:659 Failing process stderr:
ERROR FeatureFinderMetabo 0:673:659 Return code: 137
ERROR FeatureFinderMetabo 0:673:659 Execute failed: Failed to execute node FeatureFinderMetabo
ERROR Parallel Chunk End 0:609 Execute failed: Not all chunks finished - check individual chunk branches for details.

The weird part is that after the 3 branches finish successfully, if I reset the branch that fails and execute again, it finishes successfully.

Is it a memory issue? I increased Knime RAM up to 16Gb(in Knime.ini file) and keeps failing. I also increased the threads in FeatureFinderMetabo, but still nothing. Any ideas why this happens?

Hi!

Indeed this could be a memory issue. And actually increasing RAM for KNIME counter-intuitively made it worse.
The GenericKNIMENodes (e.g. used for OpenMS) will spawn command line processes
with Java’s ProcessBuilder. This reserves virtual memory in the same size of your KNIME’s JVM. If you increase the memory that KNIME uses, a) the processes will have less physical memory to work with (depending on how hungry KNIME is at the moment) and b) the processes will request even more virtual memory.

Maybe limit the number of concurrent chunks.

Best J

2 Likes

Hi!
thank you very much! That worked!
But nonetheless I removed the parallel chunk nodes and replaced them with ziploop nodes just to be sure!
Now in continuation of the workflow I get a similar error from AccurateMassSearch node that returns an exit code 134:

ERROR AcurateMassSearch Failing process stder: Terminate call after throwing an instance of ‘std::bad_alloc’
ERROR AcurateMassSearch Return code 134

Could it be a memory issue again? I have set “Knime maximum threads” to 16 and have given knime 6144m of RAM in knime.ini.