this next test was fast, sorry for the flood. > You could disable memory overcommit, that would cause things to fall > over sooner.
Again, nice idea, thanks. For reference: http://serverfault.com/questions/485798/cent-os-how-do-i-turn-off-or-reduce-memory-overcommitment-and-is-it-safe-to-do Did the above, rebooted, started chrome, it ran but did not appear visually on the screen. The process ran and allocated the memory but probably it couldnt allocate enough to run onwards. Now when i tried to run anything, all failed with the message "cannot allocate memory", and a few seconds later the X crashed again. No messages in /var/log from oom-killer (i mistyped its name in my previous email) so again it does not look like oom-killer WAS TAKING action. I'm emphasising this because i have an old message from oom-killer when it really worked: /var/log/kern.log.1:Jan 18 22:25:35 desktop kernel: [ 2027.606581] Xorg invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0 /var/log/messages.1:Jan 18 22:25:35 desktop kernel: [ 2027.606581] Xorg invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0 /var/log/syslog.1:Jan 18 22:25:35 desktop kernel: [ 2027.606581] Xorg invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0 but not this time. (my clock is set) -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20140119204324.0d588343@moli-desktop
