Date: Mon, 13 Jul 1998 16:58:19 -0700 (PDT) From: Tom Wadlow <wadlow@tw.com> To: hackers@FreeBSD.ORG Subject: Problem with swap under 3.0-980524-SNAP Message-ID: <Pine.SUN.3.95P.980713162308.8130A-100000@hotspur.tw.com>
next in thread | raw e-mail | index | archive | help
I've got an odd problem with swap space on my machine, namely, it doesn't seem to be using any. Machine: HP Kayak (Pentium II) Memory: 128M Disk: 4.3 GB SCSI /etc/fstab (with allocations shown in parens): /dev/sd0s1b none swap sw 0 0 (500MB) /dev/sd0s1a / ufs rw 1 1 (100MB) /dev/sd0s1g /export/home ufs rw 2 2 (2.1GB) /dev/sd0s1f /usr ufs rw 2 2 (1.4GB) /dev/sd0s1e /var ufs rw 2 2 (200MB) proc /proc procfs rw 0 0 Unusual things about this workstation: Two-headed display (two Matrox Millenium II's) XIG Multiheaded X11R6 instead of XFree86 (which doesn't support multihead) Kernel config: GENERIC with the mods described below FAQ: Yes, I looked, but I couldn't find anything helpful for this. Description: I'm just in the process of finalizing the setup of this system for my primary use (moving off Sun boxes for the first time in more than a decade, for whatever that's worth). Got most things working, but then ran into a problem where I ran out of pseudo-ttys. Looking through the FAQ, I figured out how to up them to the (documented) max of 64 in the kernel. Did so. Made more pseudo-ttys in /dev via MAKEDEV. Rebooted. Everything seemed to work. Ran a little script to start creating rxvts one per second to see how much stress I could put on the system. I expected this to get up into the 50s or low 60s with the kernel mods. Instead, it went up from 16 (the previous limit) to about 24, then started getting "Cannot fork" errors. I stopped the script, and everything else I tried could not fork either. Killed a few windows, enough to see that it looked like a swap problem, as in, "I am out of swap". This surprised me. So I did a "pstat -T" 136/1320 files 0M/499M swap space Hmmmm, not touching the swap space.... So I ran the test again, this time with "top" running. Turns out that the system starts hanging when mem usage gets up near the 128M border, which is the amount of RAM I have in the system. The pstat output never lifts off 0, which says to me that the system isn't using any of the 500MB of swap that I built into it. "pstat -s" says this: Device 1K-blocks Used Avail Capacity Type /dev/sd0s1b 512000 0 511872 0% Interleaved which seems to imply that the system knows about my swap space at some level as that is the correct swap volume. It is just not choosing to use that space as swap space. Any thoughts? --Tom P.S. Is the max ptys still 64 (as it says in the FAQ)? The MAKEDEV seems to imply 240 are trivially possible, with more requiring a little edit of MAKEDEV. I may try it in the kernel just to see what happens, but while I'm addressing this august body, I might as well do a little pre-emptive lobbying for more if need be. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.SUN.3.95P.980713162308.8130A-100000>