Process cassandra db configuration upon SIGUSR1 in control-node

Bug #1779186 reported by Ananth Suryanarayana on 2018-06-28
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Juniper Openstack
Status tracked in Trunk
R5.0
Fix Committed
High
Ananth Suryanarayana
Trunk
Fix Committed
High
Ananth Suryanarayana

Bug Description

During ISSU, V2 control-node needs to be notified to read the cassandra db again as rabbitmq v2 which sends update notifications would not have been up for a small time window.

Currently, DB is re-read only if the contrail-control.conf is modified (wrt cassandra db server list, user name, password, etc.). We can do the same re-read process upon SIGUSR1 signal as well. This can come in handy in case control-node is stuck ever with stale config data as well. (as a work around to restarting the control-node process itself)

Review in progress for https://review.opencontrail.org/44212
Submitter: Ananth Suryanarayana (<email address hidden>)

Review in progress for https://review.opencontrail.org/44214
Submitter: Ananth Suryanarayana (<email address hidden>)

Jeba Paulaiyan (jebap) on 2018-07-10
tags: added: contrail-control

Reviewed: https://review.opencontrail.org/44214
Committed: http://github.com/Juniper/contrail-controller/commit/a1b3a99d8988e61edb77c48e24959021655b78fe
Submitter: Zuul v3 CI (<email address hidden>)
Branch: R5.0

commit a1b3a99d8988e61edb77c48e24959021655b78fe
Author: Ananth Suryanarayana <email address hidden>
Date: Thu Jun 28 10:32:45 2018 -0700

Process cassandra db configuration upon SIGUSR1 in control-node

During ISSU, V2 control-node needs to be notified to read the cassandra
db again as rabbitmq v2 which sends update notifications would not have
been up for a small time window.

Currently, DB is re-read only if the contrail-control.conf is modified
(wrt cassandra db server list, user name, password, etc.). We can do the
same re-read process upon SIGUSR1 signal as well. This can come in handy
in case control-node is stuck ever with stale config data as well. (as a
work around to restarting the control-node process itself)

This fix was verified by systest.

TODO: Unit tests for this entire config-db re-read is still a pending
action item.

Change-Id: Ib8b9e39a572cfb688d95e81756372a1f33964d8f
Closes-Bug: 1779186

OpenContrail Admin (ci-admin-f) wrote :

Reviewed: https://review.opencontrail.org/44212
Committed: http://github.com/Juniper/contrail-controller/commit/0f4876c698c07aee44dc43197b69ac91c806db8d
Submitter: Zuul v3 CI (<email address hidden>)
Branch: master

commit 0f4876c698c07aee44dc43197b69ac91c806db8d
Author: Ananth Suryanarayana <email address hidden>
Date: Thu Jun 28 10:32:45 2018 -0700

Process cassandra db configuration upon SIGUSR1 in control-node

During ISSU, V2 control-node needs to be notified to read the cassandra
db again as rabbitmq v2 which sends update notifications would not have
been up for a small time window.

Currently, DB is re-read only if the contrail-control.conf is modified
(wrt cassandra db server list, user name, password, etc.). We can do the
same re-read process upon SIGUSR1 signal as well. This can come in handy
in case control-node is stuck ever with stale config data as well. (as a
work around to restarting the control-node process itself)

This fix was verified by systest.

TODO: Unit tests for this entire config-db re-read is still a pending
action item.

Change-Id: Ib8b9e39a572cfb688d95e81756372a1f33964d8f
Closes-Bug: 1779186

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers