Glance

Glance is listening on TCP socket before store initialization

Bug #1861723 reported by s10
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Glance
Fix Released
Undecided
Unassigned

Bug Description

When multiple backends is being used with rbd backend, Glance tries to get
fsid of the cluster using rados library.

If even one of the rbd backends is unavailable, glance-api wsgi fails to start properly,
but continue listening on the tcp socket and not responding at any request until the timeout.

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to glance (master)

Related fix proposed to branch: master
Review: https://review.opendev.org/705526

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Related fix merged to glance (master)

Reviewed: https://review.opendev.org/705526
Committed: https://git.openstack.org/cgit/openstack/glance/commit/?id=d9902c26a1f75149793b92d92968c40288f3d9ae
Submitter: Zuul
Branch: master

commit d9902c26a1f75149793b92d92968c40288f3d9ae
Author: Vlad Gusev <email address hidden>
Date: Mon Feb 3 22:19:38 2020 +0300

    Configure Glance TCP socket after the store initialization

    When multiple stores are being used with rbd backend, Glance tries
    to get fsid of the cluster using rados library.
    If even one of the rbd backends is unavailable, glance-api wsgi fails
    to start properly, but continue listening on the TCP socket and is not
    responding at any request until the timeout.

    This change moves socket configuration after the store initialization
    to make sure that failed Glance API process will not be
    considered as up by the haproxy TCP check.

    Change-Id: I4298bb835ec60b79a0010e52318ada19eb24ad80
    Related-Bug: #1861723

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to glance (stable/train)

Related fix proposed to branch: stable/train
Review: https://review.opendev.org/707653

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Related fix merged to glance (stable/train)

Reviewed: https://review.opendev.org/707653
Committed: https://git.openstack.org/cgit/openstack/glance/commit/?id=a1fd2b4de576b0bcb718a042cf257ce819543dbf
Submitter: Zuul
Branch: stable/train

commit a1fd2b4de576b0bcb718a042cf257ce819543dbf
Author: Vlad Gusev <email address hidden>
Date: Mon Feb 3 22:19:38 2020 +0300

    Configure Glance TCP socket after the store initialization

    When multiple stores are being used with rbd backend, Glance tries
    to get fsid of the cluster using rados library.
    If even one of the rbd backends is unavailable, glance-api wsgi fails
    to start properly, but continue listening on the TCP socket and is not
    responding at any request until the timeout.

    This change moves socket configuration after the store initialization
    to make sure that failed Glance API process will not be
    considered as up by the haproxy TCP check.

    Change-Id: I4298bb835ec60b79a0010e52318ada19eb24ad80
    Related-Bug: #1861723
    (cherry picked from commit d9902c26a1f75149793b92d92968c40288f3d9ae)

tags: added: in-stable-train
Abhishek Kekane (abhishek-kekane)
Changed in glance:
status: New → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

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