[rrd-users] Huge memory usage of rrdtool

Tobias Oetiker tobi at oetiker.ch
Thu Apr 22 11:51:49 CEST 2010


Hi Masahiko,

if you could provide a simple rrd file and the graphing
instructions you use (just as atext file), this would helpful
for debugging this problem.

cheers
tobi

 Today Masahiko KIMOTO wrote:

>  > > By replacing only rrdtool to 1.4.3, it seems to eat memory.
>  >
>  > even more memory ?
>
> same as 1.4.2 and it's still increasing.
>
> Test case: repeat generating a graph from a RRD per sec.
>
> Results:
> ----
> In case of 1.4.2:
> USER         PID %CPU %MEM   VSZ   RSS  TT  STAT STARTED      TIME COMMAND
> (2010/4/22 14:18:16 JST)
> kimoto     36159  1.5  0.3 55036  9800  p8  S+    2:18PM   0:00.19 /usr/local/bin/rrdtool -
> (2010/4/22 14:24:18 JST)
> kimoto     36159  3.2  0.3 56060 10284  p8  S+    2:18PM   0:12.06 /usr/local/bin/rrdtool -
> (2010/4/22 14:31:37 JST)
> kimoto     36159  3.5  0.3 56060 10672  p8  S+    2:18PM   0:26.57 /usr/local/bin/rrdtool -
> (2010/4/22 14:54:01 JST)
> kimoto     36159  2.7  0.4 57084 12012  p8  S+    2:18PM   1:09.32 /usr/local/bin/rrdtool -
> (2010/4/22 15:06:27 JST)
> kimoto     36159  2.2  0.4 58108 12760  p8  S+    2:18PM   1:27.91 /usr/local/bin/rrdtool -
> (2010/4/22 15:11:10 JST)
> kimoto     36159  2.0  0.4 59132 13084  p8  S+    2:18PM   1:34.59 /usr/local/bin/rrdtool -
> (2010/4/22 15:20:12 JST)
> kimoto     36159  2.1  0.4 59132 13648  p8  S+    2:18PM   1:47.44 /usr/local/bin/rrdtool -
>
> ----
> In case of 1.4.3:
> (2010/4/22 13:09:08 JST)
> USER         PID %CPU %MEM   VSZ   RSS  TT  STAT STARTED      TIME COMMAND
> kimoto     35967  1.2  0.3 55060  9692  p8  S+    1:09PM   0:00.20 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 13:20:30 JST)
> kimoto     35967  3.7  0.3 56084 10396  p8  S+    1:09PM   0:17.75 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 13:24:42 JST)
> kimoto     35967  3.2  0.3 56084 10696  p8  S+    1:09PM   0:23.96 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 13:29:37 JST)
> kimoto     35967  2.4  0.4 57108 10992  p8  S+    1:09PM   0:31.32 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 14:06:21 JST)
> kimoto     35967  3.2  0.4 59156 13388  p8  S+    1:09PM   1:29.32 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 14:17:44 JST)
> kimoto     35967  3.2  0.5 60180 14168  p8  S+    1:09PM   1:51.71 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
>
> In case of 1.4.3 after upgrading cairo(=1.8.8)/pango(=1.26.2)/fontconfig(=2.8.0)
> /freetype(=2.3.11):
> (2010/4/22 15:41:55 JST)
> USER         PID %CPU %MEM   VSZ   RSS  TT  STAT STARTED      TIME COMMAND
> kimoto     99172  1.0  0.3 55112 10044  p8  S+    3:41PM   0:00.14 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 15:51:10 JST)
> kimoto     99172  1.9  0.3 56136 10656  p8  S+    3:41PM   0:13.29 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 16:05:03 JST)
> kimoto     99172  2.1  0.4 57160 11552  p8  S+    3:41PM   0:32.94 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 16:14:34 JST)
> kimoto     99172  2.2  0.4 57160 12320  p8  S+    3:41PM   0:46.40 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 16:21:25 JST)
> kimoto     99172  2.1  0.4 58184 12560  p8  S+    3:41PM   0:56.08 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 16:31:28 JST)
> kimoto     99172  2.1  0.4 58184 13176  p8  S+    3:41PM   1:10.18 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 16:36:47 JST)
> kimoto     99172  2.2  0.4 59208 13552  p8  S+    3:41PM   1:17.76 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
> (2010/4/22 16:42:59 JST)
> kimoto     99172  2.1  0.4 59208 13944  p8  S+    3:41PM   1:27.06 /local/home/kimoto/source/rrdtool-1.4.3/src/.libs/rrdtool -
>
> ---------------------------------------------------------------------------
>
> We've applied change that generating graph is done by other process.
> So, now that this is not a serious problem for us.
>
> However I'm wondering other people face to this memory leaking problem or not.
>
> ----------------------------------------------------------------------
> Masahiko KIMOTO, Ph. D.
> E-mail: kimoto at ohnolab.org        URL: http://www.ohnolab.org/~kimoto
>
>

-- 
Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland
http://it.oetiker.ch tobi at oetiker.ch ++41 62 775 9902 / sb: -9900



More information about the rrd-users mailing list