[tor-bugs] #22255 [Core Tor/Tor]: Frequent OOM kills of tor process
Tor Bug Tracker & Wiki
blackhole at torproject.org
Sat Jun 3 01:12:28 UTC 2017
#22255: Frequent OOM kills of tor process
--------------------------+------------------------------------
Reporter: DeS | Owner:
Type: defect | Status: new
Priority: High | Milestone: Tor: 0.3.0.x-final
Component: Core Tor/Tor | Version:
Severity: Major | Resolution:
Keywords: | Actual Points:
Parent ID: | Points:
Reviewer: | Sponsor:
--------------------------+------------------------------------
Comment (by cypherpunks):
some observations:
minimal logged free memory was 84MB at 2017-06-02 18:59:11
https://trac.torproject.org/projects/tor/attachment/ticket/22255/mem.2.log#L2256
{{{
2017-06-02 18:59:11 3786 3702 84 0
0 4
}}}
but
https://trac.torproject.org/projects/tor/attachment/ticket/22255/PROC_MEMINFO.log
tells another story? (lower polling rate), lowest value at:
{{{
Fre Jun 2 18:45:49 CEST 2017
MemTotal: 3877580 kB
MemFree: 139124 kB
}}}
already better after that:
{{{
Fre Jun 2 18:56:41 CEST 2017
MemTotal: 3877580 kB
MemFree: 3052860 kB
}}}
> 4000MB - (300MB +400 MB) = 3300MB
you have 3786 MB memory (not 4000 MB)
I find percentage (or at least '-k') output with relevant process at the
top much more readable/useful:
{{{
smem -pr
}}}
(you might also reduce the output to relevant columns only)
{{{
-c 'command rss swap'
}}}
> How can a rising number of connections (to tor - nothing else running)
drain the 4G memory out of the machine when the tor processes itself only
use about 1G.
My feeling is that tor uses about 500MB per process when it gets killed
but uses much more before that.
Running 'smem' every 60 seconds might provide some valuable data.
--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/22255#comment:53>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
More information about the tor-bugs
mailing list