Travis / failures on ldb, tevent, etc

Ian Stakenvicius axs at
Mon Feb 13 19:32:44 UTC 2017

On 13/02/17 01:18 PM, Stefan Metzmacher wrote:
> Hi Ian,
>>>>> On a separate note, i'm also seeing the following, which looks to be
>>>>> an unrelated failure related to the cleanup of the test??
>>>> I never got it to run that far.  I agree it is odd.  Punt that one back
>>>> to the list and ask metze.
>>> It's not just my PR that does this, all open PR's from December 2016
>>> onwards seems to have this issue right now.  Will open a new thread.
>> I tracked this down to commit 278c92126, which added a bunch of 'stat'
>> calls on various dirs, including ${testbase}/tmp in cleanup() -- this
>> was meant to help debug 'flakey' tests where tmp didn't exist, but it
>> doesn't seem to make much sense to keep debugging code like this in
>> place when it definitely causes failures for all test runs when
>> they've overwise completed successfully.
>> I've got which reverts the
>> one particular stat call which seems to be causing failures on
>> --tail runs for ldb, talloc, tevent and replace.  Travis
>> CI runs should be green going foward (assuming checks do in fact pass,
>> of course)
> Can we change this to:
>  run_cmd("stat %s || true" % test_tmpdir, show=True)
> It's really not expected that this directory is missing,
> we always create it before we start with the tests and the
> cleanup() function should be the only place that removes it again.
> But it seems it's sometimes disappears earlier than expected.
> I think we should really try to understand where test_tmpdir gets removed.
> metze

Yep, switching to a non-fatal call would definitely accomplish the
same thing here.  PR updated; I've already ran it locally on ldb and
it passed as well.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: OpenPGP digital signature
URL: <>

More information about the samba-technical mailing list