[tahoe-lafs-trac-stream] [Tahoe-LAFS] #3025: Sometimes the test suite hangs in/around allmydata.test.test_runner.RunNode.test_introducer

[Tahoe-LAFS]

#3025: Sometimes the test suite hangs in/around
allmydata.test.test_runner.RunNode.test_introducer
-------------------------+-----------------------
     Reporter:  exarkun  |      Owner:
         Type:  defect   |     Status:  new
     Priority:  normal   |  Milestone:  undecided
    Component:  unknown  |    Version:  1.12.1
   Resolution:           |   Keywords:
Launchpad Bug:           |
-------------------------+-----------------------

Comment (by exarkun):

 So, my interpretation here is that "tahoe start" doesn't work in the CI
 environment because it performs standard *NIX daemonization.  This relies
 on having a correctly functioning init process which will eventually reap
 the zombie child process.  A number of our CI environments don't have a
 correctly functioning init process so the zombies linger forever.  "tahoe
 restart" and "tahoe stop" can't deal with this situation.  They wait
 forever for the zombies to get reaped.  This leads the test suite to time
 out.

 One possible fix is to give the CI environments a correctly functioning
 init process.  That's probably a good idea.

 Another possible fix is to stop using "tahoe start" in these tests and use
 the non-daemonizing "tahoe run" instead.  We know that we're trying to
 switch from "start" to "run" so this seems like a useful step in two ways.

 Probably should eventually do both but for now I'm going to do the latter.

--
Ticket URL: <https://tahoe-lafs.org/trac/tahoe-lafs/ticket/3025#comment:5>
Tahoe-LAFS <https://Tahoe-LAFS.org>
secure decentralized storage