[tor-bugs] #1789 [Tor Relay]: Wake-up from Hibernation Occurs Day 1 Each Month

[Tor Bug Tracker & Wiki]

#1789: Wake-up from Hibernation Occurs Day 1 Each Month
---------------------------------+------------------------------------------
 Reporter:  BarkerJr             |        Type:  defect   
   Status:  new                  |    Priority:  minor    
Milestone:  Tor: 0.2.2.x-final   |   Component:  Tor Relay
  Version:  Tor: 0.2.2.14-alpha  |    Keywords:           
   Parent:                       |  
---------------------------------+------------------------------------------

Comment(by mwenge):

 Replying to [ticket:1789 BarkerJr]:
 > This is happening on several relays, on both 64-bit and 32-bit CentOS
 Linux.
 >
 > Jun 01 00:00:00.981 [notice] Configured hibernation.  This interval
 began at 2010-06-01 00:00:00; the scheduled wake-up time was 2010-06-01
 00:00:00; we expect to exhaust our quota for this interval around
 2010-07-01 00:00:00; the next interval begins at 2010-07-01 00:00:00 (all
 times local)
 > Jul 01 00:00:00.388 [notice] Configured hibernation.  This interval
 began at 2010-07-01 00:00:00; the scheduled wake-up time is 2010-07-01
 04:07:55; we expect to exhaust our quota for this interval around
 2010-07-31 22:47:55; the next interval begins at 2010-08-01 00:00:00 (all
 times local)
 > Aug 01 00:00:00.152 [notice] Configured hibernation.  This interval
 began at 2010-08-01 00:00:00; the scheduled wake-up time was 2010-08-01
 00:00:00; we expect to exhaust our quota for this interval around
 2010-09-01 00:00:00; the next interval begins at 2010-09-01 00:00:00 (all
 times local)
 > Aug 01 00:00:00.184 [notice] Hibernation period ended. Resuming normal
 activity.

 I think the reason it woke up from hibernation at the start of each month
 is that it was agnostic in June and then very optimistic in July and
 August about your bandwidth usage vs your stated capacity of 200GB for the
 period. So in order to squeeze all that capacity on a relay that has quite
 low observed bandwidth it allotted an accounting period of almost the
 entire month in each case. That behaviour conforms with my understanding
 of the code.

 The real problem, it seems to me, is why it got your expected usage so
 wrong in July when it had the whole of June to learn. I also don't
 understand why it went into low bandwidth mode twice in June and July.

 > Jun 05 07:24:03.517 [notice] Bandwidth soft limit reached; commencing
 hibernation.
 > Jun 20 17:38:47.640 [notice] Bandwidth soft limit reached; commencing
 hibernation.
 > Jul 08 07:12:20.943 [notice] Bandwidth soft limit reached; commencing
 hibernation.
 > Jul 15 00:12:32.688 [notice] Bandwidth soft limit reached; commencing
 hibernation.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/1789#comment:7>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online