Date: Tue, 21 May 1996 09:38:50 +1000 From: Stephen Hocking <sysseh@devetir.qld.gov.au> To: current@freebsd.org Cc: toor@dyson.iquest.net Subject: Possible problem with new VM code? Message-ID: <199605202338.XAA13440@netfl15a.devetir.qld.gov.au>
next in thread | raw e-mail | index | archive | help
I compiled up the new VM patches (as at ctm src-cur 1804) and installed them on my home machine (486/66, 8Mb, 1542B, barracuda). Whilst running the X server & two xterms, I ran systat -vmstat in one and started doing a make all install of libc_r in the other. It seemed to be paging more heavily than I expected, so I switched WM focus (I was using 9wm as the window manager) to the window running systat. This took far longer than I expected (9wm only has about 32k of text, without counting the shared libs & my Xserver has PEX & other cruft stripped out), with a high level of disk activity being maintained. I typed :q to get out of systat, which led to an even higher level of disk activity. The system then froze. I have a kernel with DDB compiled in, but with running X, I could not switch back to the console to see where it had died, or provoke a dump to dissect later. Now I compiled all the userland binaries recently, and systat did seem to be picking up all the right stats, but I was wondering if systat could have been leaking memory at a furious rate and hence caused problems by exhausting VM, owing to some sort of mismatch in what it thought kernel structures looked like and how they actually were. I'll try recompiling everything else & get back to you. Stephen -- The views expressed above are not those of the Worker's Compensation Board of Queensland, Australia.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199605202338.XAA13440>