Some requests to Blazar hang after sending a SIGHUP to blazar-manager

[Pierre Riteau]

After sending a SIGHUP signal to blazar-manager, most (but not all) requests to Blazar don't receive any answer. It's as if RPC messaging between API and manager is partially broken.

Hint from Masahito:

09:56:37 <masahito> I guess RPCServer in blazar.utils.service should need to have SIGHUP signal handler.

Discovered while testing https://review.openstack.org/#/c/585847

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/650635

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/650635
Committed: https://git.openstack.org/cgit/openstack/blazar/commit/?id=89bfc325caa86fd0af4eaa691d04e6772d55f588
Submitter: Zuul
Branch: master

commit 89bfc325caa86fd0af4eaa691d04e6772d55f588
Author: asmita singh <email address hidden>
Date: Wed Feb 27 13:40:32 2019 +0000

    Handle SIGHUP signal for blazar-manager correctly

    After a SIGHUP signal is sent to blazar-manager, the RPC messaging
    between API and blazar-manager stops working and users start seeing
    "Internal Server Error".

    This issue is fixed in this patch, which adds a wait method to the
    RPCServer class and calls the wait method of the RPC server from
    oslo.messaging, so that all existing requests are fulfilled before
    restarting the blazar-manager service again.

    Change-Id: I8272975afeabfb74f974e9bf9a6ae9162c22a6e0
    Closes-Bug: #1807689

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/blazar 4.0.0.0rc1 release candidate.