 |
 |
Hi, in the Bari Xrootd regional redirector (recently moved to 4.1.1) we have seen a couple of problems due to the 100Hz+ rate of spawning of threads, leading to a stuck situation: at first we had a limit at 4096 threads, and we got a 150408 16:46:30 31859 XrdSched: Now have 4011 workers 150408 16:46:30 31859 XrdSched: runningredirector inq=0 150408 16:46:30 31860 XrdScheduler: Unable to create worker thread; resource temporarily unavailable problem is that it got from 50 workers to 4011 in < 1 min, which i cannot believe is due to incoming requests ... the twin redirector in pisa, still on 3.3.6, sits at 200 workers and never had a problem. they are DNS balanced so in principle should behave in the same way. today it crashed again with 8192 threads, same behavior. Anything known in 4.1.1 which can explain this? thanks a lot tom ... 150408 16:46:29 31786 XrdSched: Now have 3938 workers 150408 16:46:29 31786 XrdSched: running redirector inq=0 150408 16:46:29 31787 XrdSched: Now have 3939 workers 150408 16:46:29 31787 XrdSched: running redirector inq=0 150408 16:46:29 31788 XrdSched: Now have 3940 workers 150408 16:46:29 31788 XrdSched: running redirector inq=0 150408 16:46:29 31789 XrdSched: Now have 3941 workers ... -- Tommaso Boccali INFN Pisa ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the XROOTD-L list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1