Uploaded image for project: 'OpenShift Pipelines'
  1. OpenShift Pipelines
  2. SRVKP-2160

tekton-resource-pruner job failing when having many containers

XMLWordPrintable

    • 5
    • False
    • None
    • False
    • Hide
      * Previously, with the resource pruner, each namespace operator created a command that ran in a separate container. This design consumed too many resources in clusters with a high number of namespaces. For example, to run a single command, a cluster with 1000 namespaces produced 1000 containers in a pod.
      +
      This update fixes the issue. It passes the namespace-based configuration to the job so that all the commands run in one container in a loop. (link:https://issues.redhat.com/browse/SRVKP-2160[SRVKP-2160])
      Show
      * Previously, with the resource pruner, each namespace operator created a command that ran in a separate container. This design consumed too many resources in clusters with a high number of namespaces. For example, to run a single command, a cluster with 1000 namespaces produced 1000 containers in a pod. + This update fixes the issue. It passes the namespace-based configuration to the job so that all the commands run in one container in a loop. (link: https://issues.redhat.com/browse/SRVKP-2160 [ SRVKP-2160 ])
    • Pipelines Sprint 219

      Tekton resource pruner on cluster with 80 namespaces creates 80 containers in pod, where each runs tkn command. Few containers are failing with `Error: Get "https://172.30.0.1:443/api?timeout=32s": dial tcp 172.30.0.1:443: i/o timeout`.

       

      # oc get TektonConfig config -o jsonpath='{.spec.pruner}'
{"keep":10,"resources":["pipelinerun"],"schedule":"0/10 * * * *"}

       

      # oc logs --all-containers=true -f -n openshift-pipelines --max-log-requests=100 tekton-resource-pruner-97kx9-27488710-zxhfv 
All but 10 PipelineRuns(Completed) deleted in namespace "mkovarik"
All but 10 PipelineRuns(Completed) deleted in namespace "default" 
Error: Get "https://172.30.0.1:443/api?timeout=32s": dial tcp 172.30.0.1:443: i/o timeout 
All but 10 PipelineRuns(Completed) deleted in namespace "damoreno"
Error: Get "https://172.30.0.1:443/api?timeout=32s": dial tcp 172.30.0.1:443: i/o timeout
...

      Pipeline version: v0.28.3

            pradkuma PRADEEP KUMAR (Inactive)
            mkovarik@redhat.com Michal Kovarik
            Votes:
            0 Vote for this issue
            Watchers:
            7 Start watching this issue

              Created:
              Updated:
              Resolved: