LinHES Forums
http://forum.linhes.org/

R5B7 - memory leaks ?
http://forum.linhes.org/viewtopic.php?f=6&t=9339		 Page 1 of 1
Author:  Warped [ Tue Mar 28, 2006 1:39 pm ]
Post subject:  R5B7 - memory leaks ?
Hi *

Few weeks ago I upgrade from A22 to B7 - smoothly & without problem (minor issue was LVM migration script - it expect standard volume name vg. If you have other name - script goes wrong...)

What take my attention: during standard work (2-3 programs recorded daily) I see constant increasing allocated memory by mythfrontend process:

at start it was 178M virtual & 208M res (0 swap)
after few days it was 240M virtual, 210M res (80M swap)
after week it is 608M virt, 216M res (350M swap)

I have observation constant increase of res & swap :-(

It was not a case for A22 - so I suspect memory leaks in B7.

I'm wonder what is mem consumption in other peoples ?
Author:  mac [ Tue Mar 28, 2006 8:01 pm ]
Post subject: 
Does restarting the frontend free up the memory? You could see if there is a ticket open
for the issue at http://svn.mythtv.org/trac, if not you could open one..

I added this to my /etc/crontab to see if I am having the same issue.

========== created /root/leaking first ========

55 23 * * * root free > /root/leaking/`date +%m%d%y`
Author:  tjc [ Tue Mar 28, 2006 8:21 pm ]
Post subject:  Re: R5B7 - memory leaks ?
Warped wrote:
Few weeks ago I upgrade from A22 to B7 - smoothly & without problem (minor issue was LVM migration script - it expect standard volume name vg. If you have other name - script goes wrong...)

Duh-oh! :oops: Sigh... I should probably parse the output of the vgscan to verify the name... The only problem I foresee with that is the multiple VGs case... :? Hmmm...
Author:  ed3120 [ Wed Mar 29, 2006 9:59 am ]
Post subject: 
I never really looked before this week, (because it was only an SD system running R5A26) but last week I setup R5B7, an HD5000, and a 1080i modeline and went to a full HD system. Now I can see that my system constantly uses almost all of my 1GB of memory, which seems a little high to me.

It seems to grab all but about 15MB of memory. (TOP reports back about 15MB free.) My swap usage says 0, so apparently it hasn't maxed out the memory and hit the swap.

I very well could be wrong, but I'm also thinking memory leak.
Author:  Xsecrets [ Wed Mar 29, 2006 11:43 am ]
Post subject: 
well most any linux system at any time will be using almost all of your physical RAM because it pages to it to improve performance. If you are not using swap you are ok. The poster above however is using swap, so that is curious.
Author:  ed3120 [ Wed Mar 29, 2006 12:26 pm ]
Post subject: 
Good to know...in the Windows world it doesn't work that way, so I assumed Linux handled memory similarly.

Thanks!
Author:  mac [ Sun Apr 30, 2006 2:39 am ]
Post subject: 
I can say with some certainty that I don't have a memory leak. We tape about 4 hours per
day..

Code:

root@mythtv:~/leaking# uptime
 16:36:03 up 23 days,  1:13,  3 users,  load average: 0.00, 0.00, 0.00

======== Sun Apr 30 15:55:02 UTC 2006 =========
             total       used       free     shared    buffers     cached
Mem:        514776     508520       6256          0      22860     288376
-/+ buffers/cache:     197284     317492
Swap:       915696          0     915696
Author:  marc.aronson [ Sun Apr 30, 2006 9:01 am ]
Post subject: 
Can't say that it's a memory leak as I haven't been monitoring it over time, but my system w/ 512MB of memory shows 122MB of swap in use. I halt gdm every night at 1:15AM to prevent "mythhd kernel: NVRM: Xid (0001:00): 16, Head 00000000 Count 00067d5d" errors. Restarting gdm does not increase my swap usage. I've seen people indicate that HD systems runs better with 1GB of RAM, so I suspect my problem is just not enough RAM as opposed to a leak, but I'll check it again in a few days to see if things grow.

Marc

Code:
uptime; free -m
 07:54:42 up 3 days,  4:27,  2 users,  load average: 0.00, 0.00, 0.00
             total       used       free     shared    buffers     cached
Mem:           501        463         38          0         78        179
-/+ buffers/cache:        205        296
Swap:          815        122        693
Author:  tjc [ Sun Apr 30, 2006 9:33 am ]
Post subject: 
If there's a leak it's not obvious to me either... I've got no swap used after 8 days of uptime.
Code:
root@black2:~# free -m
             total       used       free     shared    buffers     cached
Mem:           501        411         89          0         85         46
-/+ buffers/cache:        279        221
Swap:          666          0        666
root@black2:~# uptime
 11:31:45 up 8 days, 10:58,  3 users,  load average: 0.05, 0.01, 0.00
Author:  marc.aronson [ Sun May 07, 2006 12:56 pm ]
Post subject: 
FYI -- I had 73MB of swap in use. I shut down apache (/etc/init.d/apache stop) and swap use when down to 33MB. I restarted apache and swap use stayed at 33MB. Apache is used extensively in my house for mythweb.


Marc
Author:  Liv2Cod [ Sun May 07, 2006 7:39 pm ]
Post subject: 
I have found that my HD system runs best with 1G of memory. I tried 512M at one time but found that swapping would result -- especially when the commercial flagging started up. The comm flagging seems to take a ton of memory, at least on my system.

I've been running 24/7 on my 1G machine and its very stable over time. It uses most of the memory (at least most of the time) but never seems to go into swapping.
Page 1 of 1 All times are UTC - 6 hours
Powered by phpBB® Forum Software © phpBB Group
http://www.phpbb.com/