[tahoe-dev] [tahoe-lafs] #616: bug in repairer causes sporadic hangs in unit tests

tahoe-lafs trac at allmydata.org
Wed Feb 11 16:00:32 PST 2009


#616: bug in repairer causes sporadic hangs in unit tests
---------------------------+------------------------------------------------
 Reporter:  zooko          |           Owner:       
     Type:  defect         |          Status:  new  
 Priority:  major          |       Milestone:  1.3.1
Component:  code-encoding  |         Version:  1.2.0
 Keywords:                 |   Launchpad_bug:       
---------------------------+------------------------------------------------
Changes (by warner):

  * milestone:  1.3.0 => 1.3.1


Comment:

 as mentioned in #611, we disabled the repair-from-corruption tests, and
 have only rarely seen lost-progress in the remaining repair-from-deletion
 test.

 Zooko fixed one bug in the repairer which would have caused lost-progress,
 but didn't see any other obvious ones.

 I've seen lost-progress in repair-from-deletion twice now (after zooko's
 fix), but it's pretty rare (and therefore hard to analyze). Since repair-
 from-deletion is supposed to be deterministic, the only entropy source
 remaining is the order in which download reads and upload writes are
 interleaved, which means it's going to be a long hard struggle to capture
 enough information for analysis.

 So we're going to push this one out to 1.3.1 . We'd like to have a perfect
 repairer in 1.3.0, but we also want to have a 1.3.0 soon, and a repairer
 which hangs once out of every thousand uses might be good enough for that.

-- 
Ticket URL: <http://allmydata.org/trac/tahoe/ticket/616#comment:1>
tahoe-lafs <http://allmydata.org>
secure decentralized file storage grid


More information about the tahoe-dev mailing list