Continuation after error "paired reads have different names"

[scg40]

I am using bwa-mem both on the server cluster of my university and on the galaxy platform.
When processing a sample with paired end data with mismatched reads, the program terminates on Galaxy but not on the server cluster.

How is it decided wether the program terminates after
if (strcmp(p[0]->name, p[1]->name) != 0)
err_fatal(func, "paired reads have different names: "%s", "%s"\n", p[0]->name, p[1]->name);
or just prints the message to the error output but continues?

I would really appreciate an explanation because I try to understand what causes the termination on Galaxy.
Thank you very much

[nicolo-tellini]

@scg40 are u using the same version of bwa?

bwa (0.7.17-r1188) stops running if it finds inconsistency in named reads across paired file.

I use an example, if my first read in _1 is @SRR13640876.1 also in _2 there will be a read, in the first position, named @SRR13640876.1. If this is not the case bwa returns an error at some point during the run aborting it.

I would recommend ordering the reads according the names and correct inconsistent situations.
paired reads are also expected to be the same number across paired files ( grep -c '^@' helps counting the reads ).