You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@ignite.apache.org by Shiva Kumar <sh...@gmail.com> on 2019/09/18 11:07:37 UTC

liveness and rediness probe configuration for Ignite on kubernetes

Hi all,
I have deployed Ignite on Kubernetes and configured liveness and readiness
probe like this
        readinessProbe:
          tcpSocket:
            port: 10800
          initialDelaySeconds: 10
          periodSeconds: 2
          failureThreshold: 60
        livenessProbe:
          httpGet:
           scheme: HTTP
           path: /ignite?cmd=version&user=ignite&password=ignite
           port: 8080
          initialDelaySeconds: 60
          periodSeconds: 20

where 10800 is SQL port and 8080 is rest port.

The problem I am facing here is, If pod restarts for some reason then
liveness probe is failing because during recovery of Ignite node(pod) it is
failing to respond to rest API (liveness probe) and this leads to again
restart by Kubernetes and pod going into crashLoopBackOff state.

Is there any other best way of configuring liveness probe?
during recovery of Ignite node, why it is failing to respond to a simple
rest API query ( /ignite?cmd=version ) ?

regards,
shiva

RE: liveness and rediness probe configuration for Ignite on kubernetes

Posted by "M. Ilya" <il...@outlook.com>.
Hi Shiva,

Could you please check that ignite-rest-http library is added to OPTION_LIBS environment parameter, i.e.:
      containers:
      - name: ignite
        image: apacheignite/ignite:2.6.0
        env:
        - name: OPTION_LIBS
          value: ignite-kubernetes,ignite-rest-http

If this library is present in OPTION_LIBS can you share your StatefulSet/Deployment configuration template?

Thank you,
Ilya.
________________________________
От: Shiva Kumar <sh...@gmail.com>
Отправлено: 18 сентября 2019 г. 11:07
Кому: user@ignite.apache.org <us...@ignite.apache.org>
Тема: liveness and rediness probe configuration for Ignite on kubernetes

Hi all,
I have deployed Ignite on Kubernetes and configured liveness and readiness probe like this
        readinessProbe:
          tcpSocket:
            port: 10800
          initialDelaySeconds: 10
          periodSeconds: 2
          failureThreshold: 60
        livenessProbe:
          httpGet:
           scheme: HTTP
           path: /ignite?cmd=version&user=ignite&password=ignite
           port: 8080
          initialDelaySeconds: 60
          periodSeconds: 20

where 10800 is SQL port and 8080 is rest port.

The problem I am facing here is, If pod restarts for some reason then liveness probe is failing because during recovery of Ignite node(pod) it is failing to respond to rest API (liveness probe) and this leads to again restart by Kubernetes and pod going into crashLoopBackOff state.

Is there any other best way of configuring liveness probe?
during recovery of Ignite node, why it is failing to respond to a simple rest API query ( /ignite?cmd=version ) ?

regards,
shiva