Hello Sebastian,
thank you for the detailed report. I have opened
https://phabricator.wikimedia.org/T427204 to follow up and investigate

HTH,
Filippo

On Wed, May 20, 2026 at 5:36 PM Sebastian Berlin via Cloud <
[email protected]> wrote:

> Quick update on this: I restarted the job with 2 CPUs and again it started
> right away.
>
> *Sebastian Berlin*
> Utvecklare/*Developer*
> Wikimedia Sverige (WMSE)
>
> E-post/*E-Mail*: [email protected]
> Telefon/*Phone*: (+46) 0707 - 92 03 84
>
>
> On Wed, 20 May 2026 at 16:50, Sebastian Berlin <
> [email protected]> wrote:
>
>> It looks like I ran into this again. A job has now been pending for 20
>> minutes. I tried another job with 3 CPUs earlier today and that worked
>> fine.
>>
>> When I ran `kubectl describe pods` I got a lot of info I don't
>> understand. There is however a warning:
>>
>>>   Warning  FailedScheduling  3m58s (x62 over 14m)  default-scheduler
>>>  0/81 nodes are available: 11 node(s) didn't match Pod's node
>>> affinity/selector, 3 node(s) had untolerated taint {
>>> node-role.kubernetes.io/control-plane: }, 3 node(s) had untolerated
>>> taint {toolforge.org/gateway: true}, 64 Insufficient cpu. preemption:
>>> 0/81 nodes are available: 17 Preemption is not helpful for scheduling, 64
>>> No preemption victims found for incoming pod.
>>
>>
>> I attached the full output. The first job is the one not wanting to start.
>>
>> *Sebastian Berlin*
>> Utvecklare/*Developer*
>> Wikimedia Sverige (WMSE)
>>
>> E-post/*E-Mail*: [email protected]
>> Telefon/*Phone*: (+46) 0707 - 92 03 84
>>
>>
>> On Mon, 23 Mar 2026 at 15:27, David Caro <[email protected]> wrote:
>>
>>>
>>> The workers currently have 8 CPU each, and from what I see there's more
>>> than
>>> enough cpu free to run a 2 cpu job (just did).
>>>
>>> It is true though that the more resources you request, the harder it
>>> will be
>>> to allocate (it might even fail to allocate), but I'm not seeing that the
>>> cluster was out of capacity at that moment.
>>>
>>> Next time it happens, you can also try running `kubectl describe pods`
>>> to show
>>> the events and some more details on why the job was stuck.
>>>
>>> We are working on a patch to add also more information to the state of
>>> the jobs
>>> that will help debug quota issues and similar things too, though not yet
>>> there.
>>>
>>> On 03/23 14:34, Sebastian Berlin via Cloud wrote:
>>> > Thanks. The job started when I lowered to 2 CPUs.
>>> >
>>> > This will probably have some impact on the runtime, which is a factor.
>>> I'd
>>> > still like to know if it's as simple as the extra CPU preventing it
>>> from
>>> > starting altogether. If so, does that mean the limit is set too high
>>> to in
>>> > practice?
>>> >
>>> > *Sebastian Berlin*
>>> > Utvecklare/*Developer*
>>> > Wikimedia Sverige (WMSE)
>>> >
>>> > E-post/*E-Mail*: [email protected]
>>> > Telefon/*Phone*: (+46) 0707 - 92 03 84
>>> >
>>> >
>>> > On Mon, 23 Mar 2026 at 12:44, Nicolas Vervelle <[email protected]>
>>> wrote:
>>> >
>>> > > Hi Sebastian,
>>> > >
>>> > > I also had this kind of problem, it was with jobs requesting a lot
>>> of cpu
>>> > > resources, 3 like you (which is the default maximum). When I reduced
>>> the
>>> > > CPU request, it started a lot faster.
>>> > >
>>> > > So I guess the cluster is too busy and never find a node with enough
>>> CPU
>>> > > available.
>>> > >
>>> > > Nicolas
>>> > >
>>> > > Le lun. 23 mars 2026 à 12:15, Sebastian Berlin via Cloud <
>>> > > [email protected]> a écrit :
>>> > >
>>> > >> A Toolforge job I'm running is stuck. I started it last Friday after
>>> > >> having stopped once before for now apparent reason. When I checked
>>> today I
>>> > >> got:
>>> > >>
>>> > >>> tools.commonsdb-registry@tools-bastion-15:~$ toolforge jobs show
>>> > >>> make-declarations-wlm-2013
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Job name:     | make-declarations-wlm-2013
>>> > >>>                                              |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Command:      | src/make_declaration.py -t mem:1.0Gi -t cpu:3.0
>>> -t
>>> > >>> version:0.1.11 --verbose batch:category-26152515 |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Job type:     | one-off
>>> > >>>                                               |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Image:        |
>>> tool-commonsdb-registry/tool-commonsdb-registry:latest
>>> > >>>                                              |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Port:         | none
>>> > >>>                                              |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | File log:     | yes
>>> > >>>                                               |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Output log:   |
>>> > >>> /data/project/commonsdb-registry/make-declarations-wlm-2013.out
>>> > >>>                         |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Error log:    |
>>> > >>> /data/project/commonsdb-registry/make-declarations-wlm-2013.err
>>> > >>>                         |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Emails:       | all
>>> > >>>                                               |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Resources:    | mem: 1.0Gi, cpu: 3.0
>>> > >>>                                              |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Replicas:     |
>>> > >>>                                               |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Mounts:       | all
>>> > >>>                                               |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Retry:        | no
>>> > >>>                                              |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Timeout:      | no
>>> > >>>                                              |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Health check: | none
>>> > >>>                                              |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Status:       | Running for 3d23m
>>> > >>>                                               |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>> | Hints:        | Run not attempted yet. Pod in 'Pending' phase.
>>> > >>>                                              |
>>> > >>>
>>> > >>>
>>> +---------------+-----------------------------------------------------------------------------------------------------+
>>> > >>
>>> > >>
>>> > >> As far as I can tell that means it hasn't actually started. It's
>>> still
>>> > >> waiting for something, but it's unclear what. I restarted it a
>>> couple of
>>> > >> times today so far, but it never gets past that stage.
>>> > >>
>>> > >> When I've run similar jobs in the past they sometimes take a while
>>> to
>>> > >> start. Eventually they have though, so I though maybe they were
>>> waiting for
>>> > >> resources or something.
>>> > >>
>>> > >> Can anyone help me get this started or explain if I've misunderstood
>>> > >> something?
>>> > >>
>>> > >> *Sebastian Berlin*
>>> > >> Utvecklare/*Developer*
>>> > >> Wikimedia Sverige (WMSE)
>>> > >>
>>> > >> E-post/*E-Mail*: [email protected]
>>> > >> Telefon/*Phone*: (+46) 0707 - 92 03 84
>>> > >> _______________________________________________
>>> > >> Cloud mailing list -- [email protected]
>>> > >> List information:
>>> > >>
>>> https://lists.wikimedia.org/postorius/lists/cloud.lists.wikimedia.org/
>>> > >>
>>> > >
>>>
>>> > _______________________________________________
>>> > Cloud mailing list -- [email protected]
>>> > List information:
>>> https://lists.wikimedia.org/postorius/lists/cloud.lists.wikimedia.org/
>>>
>>>
>>> --
>>> David Caro
>>> SRE - Cloud Services
>>> Wikimedia Foundation <https://wikimediafoundation.org/>
>>> PGP Signature: 7180 83A2 AC8B 314F B4CE  1171 4071 C7E1 D262 69C3
>>>
>>> "Imagine a world in which every single human being can freely share in
>>> the
>>> sum of all knowledge. That's our commitment."
>>>
>> _______________________________________________
> Cloud mailing list -- [email protected]
> List information:
> https://lists.wikimedia.org/postorius/lists/cloud.lists.wikimedia.org/
>


-- 
*Filippo Giunchedi*
Staff Site Reliability Engineer
Wikimedia Foundation <https://wikimediafoundation.org/>
_______________________________________________
Cloud mailing list -- [email protected]
List information: 
https://lists.wikimedia.org/postorius/lists/cloud.lists.wikimedia.org/

Reply via email to