Uploaded image for project: 'Couchbase Kubernetes'
  1. Couchbase Kubernetes
  2. K8S-2297

Allow Restore jobs to use node selector as per Backup

    XMLWordPrintable

Details

    • Improvement
    • Status: Resolved
    • Major
    • Resolution: Fixed
    • None
    • 2.2.1
    • None
    • 1

    Description

      Currently the CouchbaseCluster.Spec.Backup.NodeSelector only applies to backup jobs, it would be good to allow restore jobs to also be configurable in this way.

      A potential quick fix would be to reuse the same node selector for restore jobs - but does everyone want backup and restore to share node selectors (I imagine most would be ok with it)?

      Should we provide a separate field for restore (CRD update) that can default to the existing backup one? We could provide a quick fix using the backup one for now and a longer term change to provide a separate CRD update that defaults to the existing one.

      Attachments

        Issue Links

          No reviews matched the request. Check your Options in the drop-down menu of this sections header.

          Activity

            patrick.stephens Patrick Stephens (Inactive) created issue -
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Field Original Value New Value
            Rank Ranked higher
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Link This issue relates to CBSE-10394 [ CBSE-10394 ]
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Fix Version/s 2.3.0 [ 17600 ]
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Link This issue relates to K8S-2298 [ K8S-2298 ]
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Description Currently the CouchbaseCluster.Spec.Backup.NodeSelector only applies to backup jobs, it would be good to allow restore jobs to also be configurable in this way. Currently the CouchbaseCluster.Spec.Backup.NodeSelector only applies to backup jobs, it would be good to allow restore jobs to also be configurable in this way.

            A potential quick fix would be to reuse the same node selector for restore jobs - but does everyone want backup and restore to share node selectors (I imagine most would be ok with it)?

            Should we provide a separate field for restore (CRD update) that can default to the existing backup one? We could provide a quick fix using the backup one for now and a longer term change to provide a separate CRD update that defaults to the existing one.
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Component/s documentation [ 14113 ]
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Fix Version/s 2.2.1 [ 17700 ]

            Kept it simple with restore using same node selector.

            patrick.stephens Patrick Stephens (Inactive) added a comment - Kept it simple with restore using same node selector.
            roo.thorp Roo Thorp made changes -
            Link This issue relates to K8S-2301 [ K8S-2301 ]
            tin.tran Tin Tran added a comment -

            Hi Patrick Stephens
            This fix will also be included in Operator 2.2.1 right?

            Thank you Patrick.

            tin.tran Tin Tran added a comment - Hi Patrick Stephens This fix will also be included in Operator 2.2.1 right? Thank you Patrick.
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Assignee Simon Murray [ simon.murray ] Prateek Kumar [ prateek.kumar ]

            Short answer, yes it selects the node you run things on based on the conditions you specify.

            Probably best to have a look at this: https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/#nodeselector

            If there is no node matching what you specify then it cannot be scheduled until there is one.

            patrick.stephens Patrick Stephens (Inactive) added a comment - Short answer, yes it selects the node you run things on based on the conditions you specify. Probably best to have a look at this: https://kubernetes.io/docs/concepts/scheduling-eviction/assign-pod-node/#nodeselector If there is no node matching what you specify then it cannot be scheduled until there is one.
            lynn.straus Lynn Straus made changes -
            Link This issue blocks K8S-2276 [ K8S-2276 ]

            Testing covered by K8S-2301

            patrick.stephens Patrick Stephens (Inactive) added a comment - Testing covered by K8S-2301
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Resolution Fixed [ 1 ]
            Status Open [ 1 ] Resolved [ 5 ]
            patrick.stephens Patrick Stephens (Inactive) made changes -
            Fix Version/s 2.3.0 [ 17600 ]

            People

              prateek.kumar Prateek Kumar (Inactive)
              patrick.stephens Patrick Stephens (Inactive)
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                Gerrit Reviews

                  There are no open Gerrit changes

                  PagerDuty