Minor updates on Pike can fail with "socket is already closed"

Bug #1757487 reported by Jiří Stránský on 2018-03-21
This bug affects 1 person
Affects Status Importance Assigned to Milestone

Bug Description

Quite a confusing output from a failed update, probably due to Zaqar messaging issues:

 u' "stdout: Trying to pull repository ... ", ',
 u' "2018-03-19.3: Pulling from", ',
 u' "9a32f102e677: Already exists", ',
 u' "b8aa42cec17Waiting for messages on queue 'f813b1b0-41bb-411f-abb9-c1f3bd3541d6' with no timeout.
socket is already closed.
a: Already exists", ',
 u' "23a4bb7bf8d1: Already exists", ',
 u' "3f0a54d105a8: Already exists", ',
 u' "1f1ad46b4704: Already exists", ',
 u' "f51d192f486a: Pulling fs layer", ',

Honza Pokorny (hpokorny) wrote :

Where is this output coming from?

The 'socket is already' closed sounds like a UI error message.

Jiří Stránský (jistr) wrote :

Minor update CLI, `openstack overcloud update run`. The current intended fix is this backport


...to be tested.

Reviewed: https://review.openstack.org/554986
Committed: https://git.openstack.org/cgit/openstack/python-tripleoclient/commit/?id=158bcb2cd297d9f6e02bd033857d248d4fe2ab38
Submitter: Zuul
Branch: stable/pike

commit 158bcb2cd297d9f6e02bd033857d248d4fe2ab38
Author: Mathieu Bultel <email address hidden>
Date: Wed Dec 6 08:39:00 2017 +0100

    Get message from websocket instead from zaqarclient directly

    Use the websocket api to get and print the real time
    ansible execution for minor update instead of using the
    zaqarclient call directly.
    This is safer and avoid brutal claim of the messages.

    Manual conflict resolution on:

    Closes-Bug: #1757487
    Change-Id: I7e324b9e037197082c23a19b4e4b8832daaf5aee
    (cherry picked from commit eaa2e9ce75dbc107d9624a2a7325d3d8ec212a0a)
    (cherry picked from commit f31795fb7e294295d73604bbb0ab9c22103472cf)

tags: added: in-stable-pike

Change abandoned by Jiri Stransky (<email address hidden>) on branch: stable/pike
Review: https://review.openstack.org/559738
Reason: Not needed anymore, it was fixed by backporting different approach to the message fetching

Changed in tripleo:
milestone: rocky-1 → rocky-2

This issue was fixed in the openstack/python-tripleoclient 7.3.10 release.

Changed in tripleo:
milestone: rocky-2 → rocky-3
Changed in tripleo:
milestone: rocky-3 → rocky-rc1
Changed in tripleo:
status: Triaged → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers