[tor-relays] 0.2.4.17-rc on Pi, a couple weeks on

Gordon Morehouse gordon at morehouse.me
Thu Sep 19 22:39:05 UTC 2013



On Wed, 18 Sep 2013 18:28:18 -0700, Andy Isaacson <adi at hexapodia.org> wrote:

> On Wed, Sep 18, 2013 at 05:43:13PM -0700, Gordon Morehouse wrote:
> > Thanks, Roger.  I'm still not sure what finally caused the OOM-killer
> > crash this morning after almost a couple weeks (?) of uptime.  I was
> > also seeing additional clock jump messages but didn't have time to
> > diagnose it.  The Pi does not have a battery-backed RTC so it requires
> > a clock set at each start, accomplished by 'ntpdate' and kept in time
> > with, in my case, 'openntpd'.  But, 'openntpd' wasn't complaining at
> > the time of the clock jumps that were reported by Tor, and AFAIK
> > 'ntpdate' is not scheduled to run periodically, so I don't know what's
> > causing it yet.
> 
> The "Your system clock just jumped 100 seconds forward" messages are
> unlikely to be due to NTP.  Much more likely the Tor daemon was blocked
> for a significant time period, due to swapping or similar.

I figured it might have been a really bad clock, but that seemed unlikely.  I wouldn't remotely be surprised if what you are saying is true as the Pi is pretty memory-constrained.  I'll try to correlate/reproduce it.

When the Pi goes to swap, it's getting desperate - it's swapping to a Class 10 microSD card.  :/

> What does top show?  In particular the "Mem" and "Swap" lines, and the
> process line for the Tor process.  Here's a large Xeon server running 4
> Tor daemons:
> 
> top - 18:26:40 up 16 days, 18:18,  1 user,  load average: 3.79, 3.95, 3.96
> Tasks:  99 total,   4 running,  95 sleeping,   0 stopped,   0 zombie
> Cpu(s): 69.0%us,  5.1%sy,  0.0%ni, 19.4%id,  0.0%wa,  0.0%hi,  6.5%si,  0.0%st
> Mem:   8176824k total,  6361748k used,  1815076k free,    36336k buffers
> Swap:        0k total,        0k used,        0k free,   183736k cached
> 
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
> 14567 debian-t  20   0 2003m 1.8g  21m R  109 22.7   4477:00 tor
> 30171 debian-t  20   0 2390m 1.8g  20m R  100 23.6  13544:10 tor
> 18891 debian-t  20   0 1800m 1.5g  23m R   86 19.8   3134:31 tor
>  8798 debian-t  20   0  324m 149m  31m S   21  1.9  36:48.05 tor
> 
> A similar picture from your RPi might shed some light on the situation.

*Typically,* tor consumes about 25-40% of all available RAM to handle 300-800 circuits and 0.5-2Mbps of traffic (it boots with 485MB physical available after all is said and done).  It always seems to get OOM-killed when I'm sleeping.

Best,
-Gordon M.


More information about the tor-relays mailing list