Re: misc/48282: Running ATF tests on 6.1.2 in KVM VPS leaves processes running

[Andreas Gustafsson <gson%gson.org@localhost>]

The following reply was made to PR misc/48282; it has been noted by GNATS.

From: Andreas Gustafsson <gson%gson.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: gson%gson.org@localhost (Andreas Gustafsson)
Subject: Re: misc/48282: Running ATF tests on 6.1.2 in KVM VPS leaves processes 
running
Date: Mon, 7 Oct 2013 14:55:25 +0300

 I did some more debugging of this.  First of all, I tracked
 down the problem to the fs/nfs/t_rquotad test script.
 
 Instead of running the full test suite, the problem can be triggered
 manually by running:
 
   cd /usr/tests/fs/nfs/
   atf-run t_rquotad
 
 This will yield a bunch of output including
 
   tc-so:Executing command [ rump_server -lrumpvfs -lrumpnet -lrumpnet_net 
-lrumpnet_netinet -lrumpnet_shmif -lrumpfs_nfs unix://clientsock ]
   tc-so:Executing command [ rump.ifconfig shmif0 create ]
   tc-so:Executing command [ rump.ifconfig shmif0 linkstr shmbus ]
   tc-so:Executing command [ rump.ifconfig shmif0 inet 10.1.1.100 ]
   tc-so:Executing command [ mkdir /rump/mnt ]
 
 (at this point, there is a one-minute pause)
 
   tc-end: 1381146179.891394, get_nfs_be_1_user, failed, Test case timed out 
after 60 seconds
 
 Although this output may look like the "mkdir" command is timing out,
 the timeout actually happens in the following command,
 
   mount_nfs 10.1.1.1:/export /rump/mnt
 
 which ATF fails to log (that's PR 48285).
 
 If I add a "sleep 1" between the "mkdir /rump/mnt" line and the
 "mount_nfs" line, the test passes.
 
 I still don't know what causes this, but perhaps the above contains
 enough clues for someone to figure it out.
 -- 
 Andreas Gustafsson, gson%gson.org@localhost