[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Sks-devel] memory leak
From: |
Kristian Fiskerstrand |
Subject: |
Re: [Sks-devel] memory leak |
Date: |
Thu, 12 Mar 2015 16:42:32 +0100 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.5.0 |
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
On 03/12/2015 04:19 PM, Gabor Kiss wrote:
>> Last Friday I reorganized disk partitions used by SKS. At first
>> sight it was all right but now I found, that recon process
>> consumes the whole memory:
>
>> I already restarted it yesterday but today I'm out of memory
>> again. My system: Stock Debian wheezy. Package version is
>> 1.1.5-1~bpo70+1.
>
> After a few days I have more info. Look this graph memory
> consumption: http://bakacsin.ki.iif.hu/~kissg/tmp/memory-week.png
> Since 6th of March the memory footprint of SKS is quite large. As
> you can see memory blocks of the recon process are swapping out
> continously. I got alarms because host was out of swap.
>
> I don't know why recon needs many times more memory than before. I
> just moved data area to a new (maybe no so fast) disk array.
>
> Any idea?
the page size etc might have an effect, not sure what you're running
on. For comparison see munin charts from the two most active sks
instances in my cluster (in terms of serving users).
https://oc.sumptuouscapital.com/index.php/s/TXijsxeFnjtcvqa
https://oc.sumptuouscapital.com/index.php/s/sX9ZR3NCaA3plUr
I had forgotten to restart munin-node after the last perl upgrade for
the 3rd in the cluster, that handles the actual recon to external
services :| but it shows
free -m
total used free shared buffers cached
Mem: 3013 2905 107 5 210 2312
- -/+ buffers/cache: 382 2631
Swap: 511 11 500
>
> Gabor
>
- --
- ----------------------------
Kristian Fiskerstrand
Blog: http://blog.sumptuouscapital.com
Twitter: @krifisk
- ----------------------------
Public OpenPGP key 0xE3EDFAE3 at hkp://pool.sks-keyservers.net
fpr:94CB AFDD 3034 5109 5618 35AA 0B7F 8B60 E3ED FAE3
- ----------------------------
Fabricando fit faber
Practice makes perfect
-----BEGIN PGP SIGNATURE-----
iQEcBAEBCgAGBQJVAbObAAoJEP7VAChXwav6/lEH/1WBAQuytdebd+p5vnoM/BNO
1cN0inx2xNqdOptXOVRV+cpJ3Kgllv7/p4jTvKNAAn9Fd/gftBXOhQHRTRp5/jI7
Vwif9vbdDIkxvaFVwSFTunIEkJAHWUqjx0flIXL/0SBHHloV3VFTm77VkDXt3DOC
vhy111aeKw/35fGIxbifpRhu4x0dE7kEO/MiD3cRATCbhDaLWTneWWvYGfADIXAa
h6FmF80rSlfldXAsHa0Owp+eP7C1ttG/VTSNtSkmVHOoF4aGLbUUdsi4m5xF7LfG
hr6miXTw8Lh6Xup58ntyOZqMO7gkN7MthKbYBobGDYdpR66CKjXYaQSpb7gjyQY=
=ziWZ
-----END PGP SIGNATURE-----
- [Sks-devel] memory leak, Kiss Gabor (Bitman), 2015/03/10
- Re: [Sks-devel] memory leak, Gabor Kiss, 2015/03/12
- Re: [Sks-devel] memory leak,
Kristian Fiskerstrand <=
- Re: [Sks-devel] memory leak, Kim Minh Kaplan, 2015/03/13
- Re: [Sks-devel] memory leak, robert, 2015/03/13
- Re: [Sks-devel] memory leak, Gabor Kiss, 2015/03/13
- Re: [Sks-devel] memory leak, Kiss Gabor (Bitman), 2015/03/13
- Re: [Sks-devel] memory leak, Kiss Gabor (Bitman), 2015/03/14
- Re: [Sks-devel] memory leak, Gabor Kiss, 2015/03/14
- Re: [Sks-devel] memory leak, Kim Minh Kaplan, 2015/03/16
- Re: [Sks-devel] memory leak, Kiss Gabor (Bitman), 2015/03/17
- Re: [Sks-devel] memory leak, Kim Minh Kaplan, 2015/03/17
- Re: [Sks-devel] memory leak, Kiss Gabor (Bitman), 2015/03/17