Do not change the CRD but you can roll back the operator itself I believe
Gyula
On Mon, 11 Sep 2023 at 18:52, Evgeniy Lyutikov <eblyuti...@avito.ru> wrote:
> Is it safe to rollback the operator version with replace to old CRDs?
> ------------------------------
> *От:* Evgeniy Lyutikov <eblyuti...@avito.ru>
> *Отправлено:* 11 сентября 2023 г. 23:50:26
> *Кому:* Gyula Fóra
>
> *Копия:* user@flink.apache.org
> *Тема:* Re: Flink kubernets operator delete HA metadata after resuming
> from suspend
>
>
> Hi!
> No, no one could restart jobmanager,
> I monitored the pods in real time, they all deleted when suspended as
> expected.
>
>
> ------------------------------
> *От:* Gyula Fóra <gyula.f...@gmail.com>
> *Отправлено:* 11 сентября 2023 г. 20:34:52
> *Кому:* Evgeniy Lyutikov
> *Копия:* user@flink.apache.org
> *Тема:* Re: Flink kubernets operator delete HA metadata after resuming
> from suspend
>
> Hi!
>
> I could not reproduce your issue, last-state suspend/restore seems to work
> as before.
> However these 2 logs seem very suspicious:
>
> 2023-09-11 06:02:07,481 o.a.f.k.o.o.d.ApplicationObserver [INFO
> ][rec-job/rec-job] Observing JobManager deployment. Previous status: MISSING
> 2023-09-11 06:02:07,488 o.a.f.k.o.o.d.ApplicationObserver [INFO
> ][rec-job/rec-job] JobManager is being deployed
>
> Looks like after suspending (and deleting the JobManager Deployment)
> somebody restarted the JobManager manually. Is that possible?
>
> Cheers,
> Gyula
>
> On Mon, Sep 11, 2023 at 2:59 PM Evgeniy Lyutikov <eblyuti...@avito.ru>
> wrote:
>
>> Hi all!
>> After updating the operator to version 1.6.0, suspended and resuming
>> flink jobs stopped working.
>> When job resumes, the high availability metadata is removed.
>>
>> Suspend job:
>> 2023-09-11 06:01:41,548 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Event | Info | SPECCHANGED | UPGRADE
>> change(s) detected (Diff: FlinkDeploymentSpec[job.state : running ->
>> suspended]), starting reconciliation.
>> 2023-09-11 06:01:41,548 o.a.f.k.o.r.d.AbstractJobReconciler [INFO
>> ][rec-job/rec-job] Job is in running state, ready for upgrade with
>> LAST_STATE
>> 2023-09-11 06:01:41,558 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Event | Info | SUSPENDED | Suspending
>> existing deployment.
>> 2023-09-11 06:01:41,558 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Deleting cluster with Foreground propagation
>> 2023-09-11 06:01:41,558 o.a.f.k.o.s.NativeFlinkService [INFO
>> ][rec-job/rec-job] Deleting JobManager deployment while preserving HA
>> metadata.
>> 2023-09-11 06:01:41,598 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Waiting for cluster shutdown...
>> 2023-09-11 06:01:45,667 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Waiting for cluster shutdown... (5s)
>> 2023-09-11 06:01:50,730 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Waiting for cluster shutdown... (10s)
>> 2023-09-11 06:01:55,837 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Waiting for cluster shutdown... (15s)
>> 2023-09-11 06:02:00,885 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Waiting for cluster shutdown... (20s)
>> 2023-09-11 06:02:01,895 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Cluster shutdown completed.
>> 2023-09-11 06:02:01,973 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Status | Info | SUSPENDED | The resource
>> (job) has been suspended
>> 2023-09-11 06:02:01,981 o.a.f.k.o.r.d.AbstractFlinkResourceReconciler
>> [INFO ][rec-job/rec-job] Resource fully reconciled, nothing to do...
>>
>> Resume:
>> 2023-09-11 06:02:07,481 o.a.f.k.o.o.d.ApplicationObserver [INFO
>> ][rec-job/rec-job] Observing JobManager deployment. Previous status: MISSING
>> 2023-09-11 06:02:07,488 o.a.f.k.o.o.d.ApplicationObserver [INFO
>> ][rec-job/rec-job] JobManager is being deployed
>> 2023-09-11 06:02:07,563 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Status | Info | SUSPENDED | The resource
>> (job) has been suspended
>> 2023-09-11 06:02:07,576 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Event | Info | SPECCHANGED | UPGRADE
>> change(s) detected (Diff: FlinkDeploymentSpec[job.state : suspended ->
>> running]), starting reconciliation.
>> 2023-09-11 06:02:07,649 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Status | Info | UPGRADING | The resource is
>> being upgraded
>> 2023-09-11 06:02:07,649 o.a.f.k.o.r.d.ApplicationReconciler [INFO
>> ][rec-job/rec-job] Deleting deployment with terminated application before
>> new deployment
>> 2023-09-11 06:02:07,649 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Deleting cluster with Foreground propagation
>> 2023-09-11 06:02:07,649 o.a.f.k.o.s.NativeFlinkService [INFO
>> ][rec-job/rec-job] Deleting JobManager deployment and HA metadata.
>> 2023-09-11 06:02:07,691 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Waiting for cluster shutdown...
>> 2023-09-11 06:02:07,763 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Cluster shutdown completed.
>> 2023-09-11 06:02:07,763 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Deleting Kubernetes HA metadata
>> 2023-09-11 06:02:07,820 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Waiting for cluster shutdown...
>> 2023-09-11 06:02:07,831 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Cluster shutdown completed.
>> 2023-09-11 06:02:07,975 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Status | Info | UPGRADING | The resource is
>> being upgraded
>> 2023-09-11 06:02:07,987 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Event | Info | SUBMIT | Starting
>> deployment
>> 2023-09-11 06:02:07,987 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Deploying application cluster requiring last-state from
>> HA metadata
>> 2023-09-11 06:02:07,999 o.a.f.k.o.c.FlinkDeploymentController
>> [ERROR][rec-job/rec-job] Flink recovery failed
>> 2023-09-11 06:02:08,012 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Event | Warning | RESTOREFAILED | HA metadata not
>> available to restore from last state. It is possible that the job has
>> finished or terminally failed, or the configmaps have been deleted. Manual
>> restore required.
>> 2023-09-11 06:02:08,099 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Status | Error | UPGRADING |
>> {"type":"org.apache.flink.kubernetes.operator.exception.RecoveryFailureException","message":"HA
>> metadata not available to restore from last state. It is possible that the
>> job has finished or terminally failed, or the configmaps have been deleted.
>> Manual restore required.","additionalMetadata":{},"throwableList":[]}
>> 2023-09-11 06:02:08,193 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Status | Info | UPGRADING | The resource is
>> being upgraded
>> 2023-09-11 06:02:08,218 o.a.f.k.o.l.AuditUtils [INFO
>> ][rec-job/rec-job] >>> Event | Info | SUBMIT | Starting
>> deployment
>> 2023-09-11 06:02:08,218 o.a.f.k.o.s.AbstractFlinkService [INFO
>> ][rec-job/rec-job] Deploying application cluster requiring last-state from
>> HA metadata
>> 2023-09-11 06:02:08,228 o.a.f.k.o.c.FlinkDeploymentController
>> [ERROR][rec-job/rec-job] Flink recovery failed
>>
>>
>>
>>
>> * ------------------------------ *“This message contains confidential
>> information/commercial secret. If you are not the intended addressee of
>> this message you may not copy, save, print or forward it to any third party
>> and you are kindly requested to destroy this message and notify the sender
>> thereof by email.
>> Данное сообщение содержит конфиденциальную информацию/информацию,
>> являющуюся коммерческой тайной. Если Вы не являетесь надлежащим адресатом
>> данного сообщения, Вы не вправе копировать, сохранять, печатать или
>> пересылать его каким либо иным лицам. Просьба уничтожить данное сообщение и
>> уведомить об этом отправителя электронным письмом.”
>>
>
