BBPSPLIT fails cryptically with user errors

[pinin4fjords]

Description of the bug

I had a senior moment and re-submitted an rna-seq run in Seqera Platform, modifying the reference files and not using additional_fasta but forgetting to take out bbsplit_fasta_list and reset skip_bbsplit to true. This caused a failure like

Exception in thread "Thread-7" java.lang.RuntimeException: java.io.EOFException
    at fileIO.ReadWrite.readObject(ReadWrite.java:823)
    at fileIO.ReadWrite.read(ReadWrite.java:1285)
    at fileIO.LoadThread.run(LoadThread.java:30)
Caused by: java.io.EOFException
    at java.base/java.io.ObjectInputStream$PeekInputStream.readFully(ObjectInputStream.java:2926)
    at java.base/java.io.ObjectInputStream$BlockDataInputStream.readShort(ObjectInputStream.java:3421)
    at java.base/java.io.ObjectInputStream.readStreamHeader(ObjectInputStream.java:959)
    at java.base/java.io.ObjectInputStream.(ObjectInputStream.java:397)
    at fileIO.ReadWrite.readObject(ReadWrite.java:818)
    ... 2 more
Exception in thread "Thread-5" java.lang.AssertionError: null, [I@497de40d
    at align2.Block.read(Block.java:148)
    at align2.IndexMaker4$BlockMaker.makeArrays(IndexMaker4.java:140)
    at align2.IndexMaker4$BlockMaker.run(IndexMaker4.java:128)

... and the job just hung.

So despite the ultimate cause being operator error there are two issues:

• The BBMAP fails cryptically rather than checking inputs
• When a failure occurs the process seems to hang rather than finishing with a non-zero error (at least within the couple of hour timeframe I was prepared to wait).

Command used and terminal output

$ nextflow run 'https://github.com/nf-core/rnaseq'
		 -name nostalgic_pare
		 -params-file 'https://api.tower.nf/ephemeral/TEPfBKFjWQ0C9XohEadhTA.json'
		 -with-tower
		 -r dev
		 -latest

Relevant files

No response

System information

No response

[drpatelh]

Did we find a fix for this @pinin4fjords ? Saw you removed it from the milestone. Shall we push it to the next one?

[pinin4fjords]

I don't quite know what to do about it, and we're probably not going to go digging into BBMAP to fix it. I'd like to leave it open as an aid for anyone else getting the same error, but I can't see us doing anything about it any time soon, so it shouldn't be on a milestone.

[MatthiasZepper]

Which process was it that failed? The index construction or the actual BB Split mapping?

[pinin4fjords]

Pretty sure it was the actual BBSplit, but now you're asking I'm doubting myself.