Mayor Memory LEAK
my fedora core 1 is up to date right now.
just runned yum and the system is ok. i saw the glibc-2.3.2-101.1 memory leak problem, but my sistem is still having it with glibc-2.3.2-101.4. i dont know if is glibc or another component. so i only start samba (its a emulated PDC and fileserver of my office network) and twice a day RSYNC. i can manually run rsync and the script called via cron. no problems. i had to cron 3 restarts a day to keep the server healthy. its a p II 266 256 ram. when fresh booted the server runs in < 100mb mem, a couple of hours the server has ~256 mb in use. top does not show any process with more than 30 % of mem. may be 1 with 25 and the rest less than 1% Code:
|
Looks fine to me. Do a search here on "memory leak" - this question has been asked frequently :)
If you are not experiencing lots of slowdown and swapfile thrashing, you do not have any problems with memory. Looks like your swap space is barely being used at all, and as your 'free -m' command shows, you have 154M free. No worries! |
Yup, so far this looks like misinterpreting the output of "free"
Hint: the most relevant line is the one labelled" -/+ buffers/cache: " |
guys. the system HUNGS.
it keeps swapping until its death. theres no way the system can eat 256 MB doing NOTHING ! seems to be threads that use certain amount of memory that is never being released ... i have told you guys what i know so far. but is not a "reading problem". its a "out of memory" problem. Code:
Feb 2 12:51:02 fileserver kernel: Out of Memory: Killed process 1147 (xfs). |
OK, you didn't mention that the first time :) And there are a lot of posts here from people wondering why all their memory is being used up when nothing's actually wrong - it's perfectly normal for Linux to use a lot of memory if it's available; it's not normal for it to cause problems, though.
Anyway, according to the output of 'top' and 'free' that you posted, everything looks fine. 'top' really ought to show what is using memory; is the one you posted earlier from when it's running normally, or when it's about to start having problems? Does your system log show anything unusual before you start getting out of memory errors? Maybe there's a cron job or something that is starting up periodically that is causing problems. I guess my only other suggestion would be to try disabling some things temporarily to see if that helps; it might help you narrow down which one is causing trouble. |
i was launching a script via cron.
i used kcron to set it up. the script calls several times to rsync using ssh to synchronize diff dirs on 2 diff servers. this script has one line at a time to process sequentially each dir. well. i dont know why, each time it starts the system get out of memory and dies. i was wondering that was a problem with some lib or component within FEDORA. but it think it was cron running several times my script. so i changed with kcron this job from root to another user: (System crontab). it seems to be working ok now. |
Sounds like it's definitely a problem with your script then. Is this a script you wrote yourself, or something else?
|
All times are GMT -5. The time now is 07:58 AM. |