Pipeline finished successfully but with a few failed lastz jobs

[XieHongX]

Hi,

I am running some chain files, and I found that some of my jobs were successful (chain files generated) yet there were some failed last jobs in the log file, for example:

[19/3bb0ed] process > execute_jobs (436) [100%] 519 of 519, failed: 1, retrie...
Completed at: 08-Nov-2025 23:41:18
Duration : 4h 53m 29s
CPU hours : 229.8 (0.4% failed)
Succeeded : 518
Failed : 1

Will this be a problem? I am not sure if the pipeline ignores the failed jobs if there is not too many or what happened was that all jobs were successful after retrying and the pipeline simply reports the numbers of failed ones.

I understand when the input genomes are not sufficiently masked the pipeline simply breaks. That's why I am confused that a few jobs failed yet the pipeline finished successfully.

Thanks in advance,

Hongxin

[MichaelHiller]

Do you have 519 reference scaffolds? If so, it looks like 1 chaining job died, meaning you may not have chains for this scaffold. If this is a large scaffold, it could be that you need to use more memory for these jobs.

[XieHongX]

Hi Prof. Hiller

Thanks for your response! NO, I think my genomes have fewer scaffolds than 519. And I want to show a case from another query genome (same target genome):

Completed at: 14-Nov-2025 01:30:02
Duration : 1d 3h 21m 4s
CPU hours : 927.8 (39.2% failed)
Succeeded : 518
Failed : 332

For this run, there are more failed jobs but the final succeeded job number is the same as the other one I showed earlier (518), and this run indeed finished too, generating chain file at the end (although takes a lot longer). I wander the if chains were successful in this case, after all the failed retries?

[MichaelHiller]

I am pretty sure for chaining, you have num-scaffolds jobs, as all local alignments for one scaffold need to be considered together (they could all form one long chain). No idea, why you get more chaining jobs than reference scaffolds.

If 332 failed, your chains are likely incomplete. Maybe provide more memory if that is the issue?