Racon stalled after loading overlaps

[ramsayl]

I noticed the other reports of this behaviour, so hoping additional reports can help solve the problem. Although racon runs successfully the vast majority of the time on multiple different species and iterations and samples, the first illumina polishing round with one stalls out after loading sequences on Debian 3.16.7 and does not complete after days of waiting with no CPU activity (runs on other samples with similar sized input take 3-4 hours). The process is listed as sleeping. The log file shows only:

[racon::Polisher::initialize] loaded target sequences 41.898813 s [racon::Polisher::initialize] loaded sequences 5727.600880 s [racon::Polisher::initialize] loaded overlaps 5813.592557 s

Processing and mapping of reads is identical to other samples, and all were run with racon on the same machine. RAM is not an issue.

[rvaser]

Have you tried running it again to see if it will stall again? Might be a thread deadlock, not sure what to think.