ctdb porting_tests failing

[Martin Schwenke]

On Wed, 27 Jul 2016 16:00:06 +1000, Amitay Isaacs <amitay at gmail.com>
wrote:

> On Wed, Jul 27, 2016 at 3:57 PM, Martin Schwenke <martin at meltin.net> wrote:
> 
> > On Wed, 27 Jul 2016 16:20:35 +1200, Andrew Bartlett
> > <abartlet at samba.org> wrote:
> >
> > > I often seen autobuilds in the Catalyst Cloud failing with:
> > >
> > > *FAILED* tests/cunit/porting_tests_001.sh
> > >
> > > 'install' finished successfully (6.289s)
> > > tests/tool/ctdb.continue.003.sh: 23: tests/tool/ctdb.continue.003.sh:
> > > EOF: not found
> > > tests/tool/ctdb.lvs.008.sh: 16: tests/tool/ctdb.lvs.008.sh: setup_ctdb:
> > > not found
> > > tests/tool/ctdb.stop.002.sh: 23: tests/tool/ctdb.stop.002.sh: EOF: not
> > > found
> > > tests/tool/ctdb.stop.003.sh: 23: tests/tool/ctdb.stop.003.sh: EOF: not
> > > found
> > > tests/eventscripts/60.nfs.monitor.106.sh: 1109:
> > > tests/eventscripts/60.nfs.monitor.106.sh: arithmetic expression:
> > > division by zero: "1 % 0"
> > > tests/eventscripts/60.nfs.monitor.106.sh: 1109:
> > > tests/eventscripts/60.nfs.monitor.106.sh: arithmetic expression:
> > > division by zero: "2 % 0"
> > > tests/cunit/porting_tests_001.sh: 14: [: -eq: unexpected operator
> > > make: *** [autotest] Error 1
> > >
> > > Any idea what is going on here?
> > >
> > > I assume it is related to this:
> > >
> > > if [ $EUID -eq 0 ] ; then
> > >     ok "../common/system_linux.c:607 interface 'fake' not found"
> > > else
> > >     ok "../common/system_linux.c:601 failed to open raw socket"
> > > fi
> > >
> > > Is there a better way than $EUID, or is this a red herring?  (I don't
> > > know why this would be flapping however).
> >
> > Fixes attached.  These are all the superficial ones you're seeing in
> > ctdb.stderr.
> >
> > Amitay will post the real fix for the porting test failure.  I've
> > aleady seen it, so to keep things moving it is:
> >
> > Reviewed-by: Martin Schwenke <martin at meltin.net>
> >
> > :-)
> >
> > peace & happiness,
> > martin
> >
> 
> Here are 3 fixes.  First 2 fix valgrind errors and 3rd one fixes timing
> issues in porting tests.

Reviewed-by: Martin Schwenke <martin at meltin.net>

peace & happiness,
martin