On Tue, 5 Apr 2011 10:49:15 -0500 Neil Skrypuch <neil@xxxxxxxxxxxxxx> wrote: > On April 4, 2011 8:05:16 pm FUJITA Tomonori wrote: > > On Mon, 4 Apr 2011 17:44:03 +0300 > > > > Or Gerlitz <ogerlitz@xxxxxxxxxxxx> wrote: > > > FUJITA Tomonori wrote: > > > > Yeah, this patch uses 16 global threads instead of creating 4 threads > > > > per lun. Needs some cleanups though. > > > > > > Do we know why there's per thread memory consumption of 10MB? > > > > Seems that about 10MB VSZ per pthread is normal. Try some experiment > > with simple pthread code. > > FWIW, VSZ is not a good measure of memory usage (that memory is not > necessarily in use). RES is what you want to look at. Yeah, true. > I have a process here with 5000 threads, the VSZ is 78.6g, but RES is only > 284m. Needless to say, this machine doesn't have anywhere close to 78G of RAM > + swap, but runs this process just fine. > > To the original poster: more likely you ran out of file descriptors and the > error message is just deceptive. The default per process fd limit is 1024 (2 > fds per thread = ~500). Try ulimit -n 20000, then restart tgtd in the same > shell. Hmm, I think that tgtd increases the fd limit internally. -- To unsubscribe from this list: send the line "unsubscribe stgt" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
