tripleo-ci promotion failing on "pull ppc64le tagged containers"

Bug #1861342 reported by wes hayutin
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
tripleo
Fix Released
Critical
Marios Andreou

Bug Description

2020-01-29 20:55:12,019 p=30334 u=centos | TASK [containers-promote : Pull ppc64le tagged container images from trunk.registry.rdoproject.org registry] ***
2020-01-29 20:56:06,452 p=30334 u=centos | failed: [localhost] (item=aodh-api) => {"ansible_index_var": "index", "ansible_loop_var": "item", "changed": false, "index": 0, "item": "aodh-api",
 "msg": "Error pulling trunk.registry.rdoproject.org/tripleomaster/centos-binary-aodh-api - code: None message: unexpected EOF"}
2020-01-29 20:57:00,705 p=30334 u=centos | failed: [localhost] (item=aodh-evaluator) => {"ansible_index_var": "index", "ansible_loop_var": "item", "changed": false, "index": 1, "item": "aodh
-evaluator", "msg": "Error pulling trunk.registry.rdoproject.org/tripleomaster/centos-binary-aodh-evaluator - code: None message: unexpected EOF"}
2020-01-29 20:57:54,371 p=30334 u=centos | failed: [localhost] (item=aodh-listener) => {"ansible_index_var": "index", "ansible_loop_var": "item", "changed": false, "index": 2, "item": "aodh-
listener", "msg": "Error pulling trunk.registry.rdoproject.org/tripleomaster/centos-binary-aodh-listener - code: None message: unexpected EOF"}
2020-01-29 20:58:47,712 p=30334 u=centos | failed: [localhost] (item=aodh-notifier) => {"ansible_index_var": "index", "ansible_loop_var": "item", "changed": false, "index": 3, "item": "aodh-
notifier", "msg": "Error pulling trunk.registry.rdoproject.org/tripleomaster/centos-binary-aodh-notifier - code: None message: unexpected EOF"}
2020-01-29 20:59:40,940 p=30334 u=centos | failed: [localhost] (item=barbican-api) => {"ansible_index_var": "index", "ansible_loop_var": "item", "changed": false, "index": 4, "item": "barbic
an-api", "msg": "Error pulling trunk.registry.rdoproject.org/tripleomaster/centos-binary-barbican-api - code: None message: unexpected EOF"}
^[OD2020-01-29 21:00:35,143 p=30334 u=centos | failed: [localhost] (item=barbican-keystone-listener) => {"ansible_index_var": "index", "ansible_loop_var": "item", "changed": false, "index":
5, "item": "barbican-keystone-listener", "msg": "Error pulling trunk.registry.rdoproject.org/tripleomaster/centos-binary-barbican-keystone-listener - code: None message: unexpected EOF"}
2020-01-29 21:01:35,377 p=30334 u=centos | failed: [localhost] (item=barbican-worker) => {"ansible_index_var": "index", "ansible_loop_var": "item", "changed": false, "index": 6, "item": "bar
bican-worker", "msg": "Error pulling trunk.registry.rdoproject.org/tripleomaster/centos-binary-barbican-worker - code: None message: unexpected EOF"}

Looking at the registry.. it looks like we have a match on:

 "name": "centos-binary-aodh-api:27e3154390c2690006d9fcc0656e923d9b812b67_1c9b2be6"
 "name": "centos-binary-aodh-api:27e3154390c2690006d9fcc0656e923d9b812b67_1c9b2be6_ppc64le",

    {
      "metadata": {
        "name": "centos-binary-aodh-api:27e3154390c2690006d9fcc0656e923d9b812b67_1c9b2be6",
        "namespace": "tripleomaster",
        "selfLink": "/oapi/v1/namespaces/tripleomaster/imagestreamtags/centos-binary-aodh-api%3A27e3154390c2690006d9fcc0656e923d9b812b67_1c9b2be6",
        "uid": "62788320-3cf5-11e8-93c1-fa163e7f8b33",
        "resourceVersion": "104247530",
        "creationTimestamp": "2020-01-29T08:55:14Z"
      },
      "tag": null,
      "generation": 1,
      "lookupPolicy": {
        "local": false
      },
      "image": {
        "metadata": {
          "name": "sha256:73323084422753b8983d7b2cee42590e7f7031249973e991f53038a6244bdedf",
          "creationTimestamp": null
        },
        "dockerImageReference": "docker-registry.default.svc:5000/tripleomaster/centos-binary-aodh-api@sha256:73323084422753b8983d7b2cee42590e7f7031249973e991f53038a6244bdedf",
        "dockerImageMetadata": {
          "kind": "DockerImage",
          "apiVersion": "1.0",
          "Id": "",
          "Created": null,
          "ContainerConfig": {}
        },
        "dockerImageMetadataVersion": "1.0",
        "dockerImageLayers": null
      }
    },
    {
      "metadata": {
        "name": "centos-binary-aodh-api:27e3154390c2690006d9fcc0656e923d9b812b67_1c9b2be6_ppc64le",
        "namespace": "tripleomaster",
        "selfLink": "/oapi/v1/namespaces/tripleomaster/imagestreamtags/centos-binary-aodh-api%3A27e3154390c2690006d9fcc0656e923d9b812b67_1c9b2be6_ppc64le",
        "uid": "62788320-3cf5-11e8-93c1-fa163e7f8b33",
        "resourceVersion": "104247530",
        "creationTimestamp": "2020-01-29T14:53:45Z"
      },
      "tag": null,
      "generation": 1,
      "lookupPolicy": {
        "local": false
      },

Code is here:

https://github.com/rdo-infra/ci-config/blob/79bcc9c64b82f1c6806139118e5b9a3663dcdb76/ci-scripts/container-push/roles/containers-promote/tasks/manifest-push.yml#L40-L49

Tags: alert ci
Revision history for this message
Marios Andreou (marios-b) wrote :

hmm not sure if there is something wrong or we need some kind of special handline for the ppc tagged containers.

i can reproduce this using just podman... i went to [1] and found centos-binary-aodh-api:03e7a3a58585eb6d751cb6ea765973241f12f2f9_6e3b098e_ppc64le so its definitely there

attaching the trace here but basically pulling _x86_64 is OK but pulling _ppc64le gives the error:

podman pull trunk.registry.rdoproject.org/tripleomaster/centos-binary-aodh-api:03e7a3a58585eb6d751cb6ea765973241f12f2f9_6e3b098e_ppc64le
==> Copying blob 1373db2e09f7 done
  unexpected EOF
Error: error pulling image "trunk.registry.rdoproject.org/tripleomaster/centos

BUT
podman pull trunk.registry.rdoproject.org/tripleomaster/centos-binary-aodh-api:03e7a3a58585eb6d751cb6ea765973241f12f2f9_6e3b098e_x86_64
Trying to pull trunk.registry.rdoproject.org/tripleomaster/centos-binary-aodh-api:03e7a3a58585eb6d751cb6ea765973241f12f2f9_6e3b098e_x86_64
==> Writing manifest to image destination
Storing signatures
bf4f198f197705611482a8a3dfc9acc1da7d914492a3a51335aa1fb27737bc4d
[marios@story1501rhel8containers ~]$

Revision history for this message
Marios Andreou (marios-b) wrote :
Revision history for this message
wes hayutin (weshayutin) wrote :

 I think adding another test or two two to molecule may come out of this. Most importantly.. the promoter code should be a little more rigorous in checking the ppc containers and manifest but also fall back to just an x86_64 promotion if another arch fails.

Changed in tripleo:
assignee: nobody → Gabriele Cerami (gcerami)
Revision history for this message
Marios Andreou (marios-b) wrote :

ci-config master: Handle errors pulling ppc64le containers when creating manifests https://review.rdoproject.org/r/24771

Changed in tripleo:
assignee: Gabriele Cerami (gcerami) → Marios Andreou (marios-b)
Revision history for this message
wes hayutin (weshayutin) wrote :
Changed in tripleo:
status: Triaged → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.