On 07/30/2018 01:49 PM, Miro Hrončok wrote:
On 30.7.2018 12:42, Mikolaj Izdebski wrote:
> On 07/30/2018 11:19 AM, Miro Hrončok wrote:
>> Hi,
>>
>> I've filled this bugzilla: [0] and I clearly remember seeing a failed
>> build in koschei - I've even linked it in the BZ [1].
>>
>> Yet the URL now returns 404.
>>
>> Last build in koschei for python3 is 3 days older than the bugreport.
>>
>> What happens? Are certain builds deleted for various reasons? Or are we
>> loosing some data by accident?
>
> The corresponding task is
>
https://koji.fedoraproject.org/koji/taskinfo?taskID=28608653
>
> This task was cancelled by Koschei itself and removed from database
> because it was running longer than 48 hours. Koschei does this to
> prevent stalled builds from blocking build queue and wasting resources.
I would found it less confusiong if they are kept in the DB but marked
them "stalled -> canceled"? For example here at least one architecture
failed and even the fact that the build was stalled is interesting.
Agreed. Can you file upstream issue at [1] so that we can track it?
[1]
https://github.com/msimacek/koschei/issues
--
Mikolaj Izdebski
Senior Software Engineer, Red Hat
IRC: mizdebsk