Fuel for OpenStack

Openstack configuration took more that 15 min for one node

Bug #1572919 reported by Ksenia Svechnikova on 2016-04-21

This bug report is a duplicate of: Bug #1567424: Redeploy takes too much time. Edit Remove

This bug affects 1 person

	Status	Importance	Assigned to	Milestone
Fuel for OpenStack	Confirmed	Critical	Krzysztof Szukiełojć	Fuel for OpenStack 10.0
Mitaka	Confirmed	Critical	Krzysztof Szukiełojć	Fuel for OpenStack 9.0
Newton	Confirmed	Critical	Krzysztof Szukiełojć	Fuel for OpenStack 10.0

Bug Description

Detailed bug description:
ISO fuel-9.0-225, version - http://paste.openstack.org/show/494939/ with fix for https://bugs.launchpad.net/fuel/+bug/1571542

Steps to reproduce:

1. Run fuel openstack-config --execute --env 1 --node 1/role/etc

Expected results:

New deployment task finished, config was applied in 600 sec

Actual result:
It took more that 1200 sec, it lead to fails in autotests

Reproducibility:
100% with config for different services

Impact:
All tests for service reconfiguration fails in our swarm

See original description

Tags:

Revision history for this message

Ksenia Svechnikova (kdemina) wrote on 2016-04-21:

fail_error_reconfigure_ml2_vlan_range-fuel-snapshot-2016-04-20_23-24-18.tar.xz Edit (52.6 MiB, application/octet-stream)

Changed in fuel:
assignee:	nobody → Fuel Mixed (fuel-mixed-team)
importance:	Undecided → Critical
milestone:	none → 9.0
description:	updated

Ksenia Svechnikova (kdemina) on 2016-04-21

description:

updated

Ksenia Svechnikova (kdemina) on 2016-04-21

summary:	- Openstack configuration put node in Error state + Openstack configuration took more that 15 min for onr node
summary:	- Openstack configuration took more that 15 min for onr node + Openstack configuration took more that 15 min for one node

Ksenia Svechnikova (kdemina) on 2016-04-21

tags:

added: feature-openstack-config-change swarm-blocker

Dmitry Klenov (dklenov) on 2016-04-21

tags:

added: area-library

Revision history for this message

Bartosz Kupidura (zynzel) wrote on 2016-04-21:

This is caused because nailgun executes not only needed tasks (with refresh_on property) but probably all tasks.

Ksenia Svechnikova (kdemina) on 2016-04-21

description:

updated

OpenStack Infra (hudson-openstack) on 2016-04-22

Changed in fuel:
assignee:	Bartosz Kupidura (zynzel) → Stanislaw Bogatkin (sbogatkin)
status:	Confirmed → In Progress

Revision history for this message

Bartosz Kupidura (zynzel) wrote on 2016-04-22:

Currently refresh_on field is not used by nailgun. This was changed due to LCM support in 9.0

Time of serialization/execution will be improved by:
https://review.openstack.org/#/c/307824/
https://review.openstack.org/#/c/307740/

In meantime we should increate timeout in QA tests, not to block swarm.

Revision history for this message

Ksenia Svechnikova (kdemina) wrote on 2016-04-22:

With what value do you want to fix QA tests? It's not a proper solution in increase timeout in 2 times, but we can add such commit for muster in case you will suggest the proper value for cluster, role and node execution timeout

Revision history for this message

Ksenia Svechnikova (kdemina) wrote on 2016-04-25:

Looks like it's duplicate of https://bugs.launchpad.net/fuel/+bug/1567424