Fix replication tests...

[Garming Sam]

I just saw one of these come back on the flakey builds again, and it
looks like it returns NT_STATUS_DS_BUSY, which is probably because it's
already in a run of the python KCC. It seems like an unlikely thing to
happen, although I guess it is spurious and probably will happen given
enough runs, but I'd like to see how long a periodic run of the KCC
actually takes. Otherwise, it seems like the client should plausibly be
async and have a proper timeout.


Cheers,

Garming

On 10/08/16 00:33, Stefan Metzmacher wrote:
> I had two failures, the ctdb one in 54_transaction_loop_recovery.sh
> and the followin I've never seen before:
>
> ERROR(runtime): DsExecuteKCC failed - (-1073741643, '{Device Timeout}
> The specified I/O operation on %hs was not completed before the time-out
> period expired.')
> Failed to exec kcc on remote DC
> SOCKET_WRAPPER_DEFAULT_IFACE="33"
> KRB5_CONFIG="/memdisk/metze/W/b196036/samba/bin/ab/promoted_dc/etc/krb5.conf"
> ./bin/samba-tool drs kcc localdc --configfile=/me
> mdisk/metze/W/b196036/samba/bin/ab/promoted_dc/etc/smb.conf
> -UAdministrator%locDCpass1 at
> /memdisk/metze/W/b196036/samba/selftest/target/Samba4.pm line 2212.
> failed to start up environment 'promoted_dc' at
> /memdisk/metze/W/b196036/samba/selftest/target/Samba.pm line 49.
> samba can't start up known environment 'promoted_dc' at
> /memdisk/metze/W/b196036/samba/selftest/selftest.pl line 901.
> Traceback (most recent call last):
>   File "/memdisk/metze/W/b196036/samba/selftest/filter-subunit", line
> 91, in <module>
>     ret = subunithelper.parse_results(msg_ops, statistics, sys.stdin)
>   File "/memdisk/metze/W/b196036/samba/selftest/subunithelper.py", line
> 64, in parse_results
>     msg_ops.time(dt)
>   File "/memdisk/metze/W/b196036/samba/selftest/subunithelper.py", line
> 266, in time
>     self._ops.time(time)
>   File "bin/python/samba/subunit/run.py", line 189, in time
>     time.second, time.microsecond))
> IOError: [Errno 32] Broken pipe
>
> This is a timeout but not (at least directly) related to one I've fixed now.
>
> The loop is still running more autobuilds with it.
>
> metze
>> Am 08.08.2016 um 14:26 schrieb Stefan Metzmacher:
>>> Am 05.08.2016 um 22:49 schrieb Jeremy Allison:
>>>> On Fri, Aug 05, 2016 at 10:16:42PM +0200, Stefan Metzmacher wrote:
>>>>> Am 05.08.2016 um 17:36 schrieb Jeremy Allison:
>>>>>> On Thu, Aug 04, 2016 at 11:26:46PM +0200, Stefan Metzmacher wrote:
>>>>>>> I'm currently running some autobuilds with the following
>>>>>>> patchset...
>>>>>>>
>>>>>>> metze
>>>>>> Metze - autobuilds are still failing and it's been
>>>>>> a week without real progress.
>>>>>>
>>>>>> Did your tests help at all ?
>>>>>>
>>>>>> I'm planning to push the patch to skip the broken
>>>>>> tests today unless I hear any different.
>>>>> I got distracted a bit today.
>>>>>
>>>>> I'm currently running autobuilds with the attached patches.
>>>> Oh - that looks like a good set of changes that should
>>>> fix the underlying issue. Let me know if you want me to
>>>> review.
>>> I've a new extended patchset, that also fixed a 3rd bug
>>> related to the prefix map and resolving a working schema
>>> when we try to replicated the schema partition itself.
>>>
>>> I'm running autobuilds with it.
>>> A subset already passed a few times.
>>>
>>> metze
>>>
>>>
>>>


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 490 bytes
Desc: OpenPGP digital signature
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20160819/db88e69e/signature.sig>