Pavel, It was supposed to only use 32. Thanks for reporting this. I'll look into that for you and try to get it fixed. Just out of curiosity, are you running this on a machine with some sort of RAIDed disk drives. If you only have one drive the paralloc doesn't seem to really help anyway, as the single disk is the bottleneck.
Drew Pavel Filipensky wrote: > Hi, > > I am running Filebench 1.2.4 on s10 x86. I have set the fileset to > contain 1.000.000 files, with options prealloc=100,reuse,paralloc. > During creation phase, I see that the go_filebench process is not > destroying threads - so far there is almost million lwps: > > > # ps -L -p 2746|wc -l > 741525 > > > # ps -L -p 2746|head > PID LWP TTY LTIME CMD > 2746 1 pts/1 4:03 go_fileb > 2746 2 pts/1 0:01 go_fileb > 2746 3 pts/1 0:00 <defunct> > 2746 4 pts/1 0:00 <defunct> > 2746 5 pts/1 0:00 <defunct> > 2746 6 pts/1 0:00 <defunct> > 2746 7 pts/1 0:00 <defunct> > 2746 8 pts/1 0:00 <defunct> > 2746 9 pts/1 0:00 <defunct> > > # ps -L -p 2746|tail > 2746 741473 pts/1 0:00 <defunct> > 2746 741474 pts/1 0:00 go_fileb > 2746 741475 pts/1 0:00 go_fileb > 2746 741476 pts/1 0:00 go_fileb > 2746 741477 pts/1 0:00 go_fileb > 2746 741478 pts/1 0:00 go_fileb > 2746 741479 pts/1 0:00 go_fileb > 2746 741480 pts/1 0:00 go_fileb > 2746 741481 pts/1 0:00 go_fileb > 2746 741482 pts/1 0:00 go_fileb > > Can this be fixed to save the resources? > > Thanks, > Pavel > > _______________________________________________ > perf-discuss mailing list > perf-discuss@opensolaris.org > _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org
