On Tue, Jun 24, 2008 at 1:11 PM, Alan Cox <alan@xxxxxxxxxxxxxxxxxxx> wrote: >> When the system hangs, the mouse cursor will continue to move, but >> it is very jumpy and sluggish. But otherwise the system is completely >> unresponsive (not just slow). > > That sounds like it suddenly ran out of memory. > > Try > > echo "2" >/proc/sys/vm/overcommit_memory > echo "80" >/proc/sys/vm/overcommit_ration > > and see if instead you suddenly see lots of out of memory errors and > applications going away BTW, that's _ratio, not _ration. Prior to this, those parameters were 0 and 50 respectively. Ah, yes, as soon as I ran the first command, my X11 session immediately aborted and I got sent back to the gdm login. Although my memory use should have been pretty small at the time (I had just rebooted so only a terminal window and a single Firefox tab were running). > Does anything make /var/log/messages ? Prior to changing the kernel flags above, no, nothing ever showed up in syslog. But after this I saw something like this right at the point of running the first command above: Jun 24 14:03:44 deron kernel: [drm] Num pipes: 1 Jun 24 14:03:44 deron udevd[545]: udev_event_run: fork of child failed: Cannot allocate memory Jun 24 14:03:44 deron udevd[545]: udev_event_run: fork of child failed: Cannot allocate memory Jun 24 14:03:45 deron restorecond: Will not restore a file with more than one hard link (/etc/resolv.conf) Cannot allocate memory Jun 24 14:03:45 deron kernel: tomboy[2973]: segfault at 8968548 ip 08968548 sp bfaeab7c error 4 Jun 24 14:03:45 deron acpid: client connected from 4262[0:0] Jun 24 14:03:45 deron bonobo-activation-server (dem-4259): could not associate with desktop session: Failed to connect to socket /t mp/dbus-vx4vmi4BAf: Connection refused Jun 24 14:03:46 deron kernel: agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0. Jun 24 14:03:46 deron kernel: agpgart: Putting AGP V3 device at 0000:00:00.0 into 8x mode Jun 24 14:03:46 deron kernel: agpgart: Putting AGP V3 device at 0000:01:00.0 into 8x mode Jun 24 14:03:47 deron kernel: [drm] Setting GART location based on new memory map Jun 24 14:03:47 deron kernel: [drm] Loading R300 Microcode Jun 24 14:03:47 deron kernel: [drm] Num pipes: 1 Jun 24 14:03:47 deron kernel: [drm] writeback test succeeded in 1 usecs Other than the udev errors, I don't know what to make of the others. >> This system is a Dell Optiplex GX270 (4-5 yrs old) with a Pentium 4 3GHz cpu >> with hyperthreading enabled, 1 MB ram. I am using LVM and LUKS encryption >> on the swap and /home. > > 1MB or 1GB. 1MB might be a bit slow so I suspect the latter ;) Yes, a typo, 1 GiB physical memory. > I wonder if LUKS + swap might be the first suspect Well, I looked a little closer and it may be my fault. The LVM I have swap in was only 32 MB in size, not the 32 GB I had intended! So my swap is way smaller than my physical memory. Would that excessively small swap space size had resulted in the abrupt hanging behavior I had witnessed? I'm going to try to resize things and get my swap back up to 32 GB. A little tricky due to LUKS being in the mix, but I should be able to do it. >> The only hardware change in the system was replacing my old PCI graphics >> card (Matrox Millenium II) with a brand new ATI Stealth X1050 AGP. This was >> necessary because the F9 graphical install was unusable; as the old card only >> had VGA output and the installer's X11 could not properly auto-detect the LCD >> widescreen monitor and was trying to drive it at too high a frequency. I needed >> to have DVI output for the F9 graphical install to work; so I had to get a new >> video card. > > Which video driver are you using ? The Fedora supplied driver, not the ATI add-on. I'm can't remember how to figure out which X driver is being used anymore. Thanks Alan -- Deron Meranda -- fedora-list mailing list fedora-list@xxxxxxxxxx To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
