The frontline support team will be unavailable to answer questions on April 15th and 17th 2019. We will be back soon after. Thank you for your patience and we apologize for any inconvenience!
I am having some issues with Cromwell performance. I was hoping for some guidance from the team.
I have cromwell running as a server, with mysql configured for caching. I have run about 4 workflows (~16 tasks) against this server/db. As I submit a job I see cpu usage spike to 100% for 2 threads of cromwell (using htop). The cpu usage eventually settles down, ~20 minutes into the job.
PID USER VIRT RES SHR S CPU% MEM% TIME+ Command
2338 evaben 11.7G 1440M 13644 S 101. 4.5 1h03:58 java -Dconfig.file=/home/evaben/cromwell/ ...
27815 evaben 11.7G 1440M 13644 R 100. 4.5 12:32.44 java -Dconfig.file=/home/evaben/cromwell/ ...
I have also attached a log, showing that 10 minutes was spent for a 'BwaMem' task to find a Cache Hit. I do not know the full mechanism of the cache. Could it be checking the md5 of my input files? Probably my two problems are related, the high cpu and the slow cache.
Please advise if these are expected times, or potentially some problem with my setup or a bug. I am using release 31.