Commit Graph

176 Commits (657e6c8c3b5bafb96d4179d596de0352479e2a64)

Author SHA1 Message Date
Andrew Schoen 5622c94e8b rolling-update: do not use upstart to stop mons when using systemd
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2017-02-21 12:31:26 -06:00
Sébastien Han 8f94bfb498 rolling-update: detect init system properly
Simply use the ansible_service_mgr fact.

Closes: #1286

Signed-off-by: Sébastien Han <seb@redhat.com>
2017-02-08 08:52:05 +01:00
Guillaume Abrioux 76ddcbc271 Remove support of releases prior to Jewel.
According to #1216, we need to simply the code by removing the
support of anything before Jewel.

Signed-off-by: Guillaume Abrioux <gabrioux@redhat.com>
2017-01-31 11:00:54 +01:00
Sébastien Han 829e2b6598 Merge pull request #1077 from font/rolling_update
Support containerized rolling update
2016-11-22 16:56:46 +01:00
Sébastien Han 38e846e542 rolling_update: clarify "serial" usage
Prior to this commit the serial variable was poorly documented. Now we
are making clear that this value should be left untouched as the rolling
update mechanism should happen serially.

Solves: bz-1396742
Signed-off-by: Sébastien Han <seb@redhat.com>
2016-11-21 14:42:46 +01:00
Ivan Font 255e816e28 Rolling update changes for containerized deployments
Separate out systemd restart tasks for containerized and
non-containerized deployments

Signed-off-by: Ivan Font <ifont@redhat.com>
2016-11-17 11:25:25 -08:00
Ivan Font e72f08080d Warn user when upgrading cluster with only one mon
Signed-off-by: Ivan Font <ifont@redhat.com>
2016-11-17 11:25:25 -08:00
Ivan Font 3ff17f1c8f Support containerized rolling update
- Update rolling update playbook to support containerized deployments
  for mons, osds, mdss, and rgws
- Skip checking if existing cluster is running when performing a rolling
  update
- Fixed bug where we were failing to start the mds container because it
  was missing the admin keyring. The admin keyring was missing because
  it was not being pushed from the mon host to the ansible host due to
  the keyring not being available before running the copy_configs.yml
  task include file. Now we forcefully wait for the admin keyring to be
  generated before continuing with the copy_configs.yml task include file
- Skip pre_requisite.yml when running on atomic host. This technically
  no longer requires specifying to skip tasks containing the with_pkg tag
- Add missing variables to all.docker.sample
- Misc. cleanup

Signed-off-by: Ivan Font <ifont@redhat.com>
2016-11-17 11:25:25 -08:00
Alfredo Deza 60ce2311b8 rolling_update: bump retries for osd_check/retries to 20 minutes
Signed-off-by: Alfredo Deza <adeza@redhat.com>

Resolves: rhbz#1395073
2016-11-17 10:43:58 -05:00
Sébastien Han 81a72cb85d Merge pull request #1068 from ceph/v2.2
moving to ansible v2.2 compatibility
2016-11-16 16:33:40 +01:00
Andrew Schoen 5f44b118b8 rolling update: stop RGWs before upgrade and start afterwards
Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Resolves: rhbz#1394929
2016-11-14 14:47:12 -06:00
Andrew Schoen ded9d9dfd3 rolling update: stop MDSs before upgrading and start afterwards
Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Resolves: rhbz#1394929
2016-11-14 14:47:12 -06:00
Andrew Schoen 5429c5f8c5 rolling update: stop MONs before upgrading and start afterwards
Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Resolves: rhbz#1394929
2016-11-14 14:47:12 -06:00
Andrew Schoen 66f09bdac4 rolling update: stop OSDs before upgrading
This avoids a bug where OSDs are sometimes restarted twice on
upgrades which leaves the OSD process running but not marked up.

See:

https://bugzilla.redhat.com/show_bug.cgi?id=1394928
https://bugzilla.redhat.com/show_bug.cgi?id=1391675
https://bugzilla.redhat.com/show_bug.cgi?id=1394929

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Resolves: rhbz#1394929
2016-11-14 14:46:58 -06:00
Sébastien Han 991341f525 rolling_update: add variable to upgrade ceph
My stupid self removed this crucial variable here: 217ce3ca thinking it
was another hard coded variable import where this is actually the
trigger for the upgrade.

Closes: #1071

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-11-04 17:31:02 +01:00
Sébastien Han a2fcd222d2 moving to ansible v2.2 compatibility
Signed-off-by: Sébastien Han <seb@redhat.com>
Co-Authored-By: Julien Francoz julien@francoz.net
2016-11-04 10:09:38 +01:00
Andrew Schoen 8262ce5e40 rolling update: fix restarts of radosgw
Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Resolves: rhbz#1391675
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-11-03 14:36:42 -05:00
Eduard Egorov 645b5efebf Fix hard-coded host group names in include tasks for group variables' file paths.
Signed-off-by: Eduard Egorov <eduard.egorov@icl-services.com>
2016-11-01 12:21:40 +00:00
Andrew Schoen 0897c965ff rolling_update: define mon_group_name when upgrading the mons
see: https://bugzilla.redhat.com/show_bug.cgi?id=1389456

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

Resolves: rhbz#1389456
2016-10-27 14:17:56 -05:00
Sébastien Han b0989c700f rolling_update: fix wrong indent
Fixing: https://bugzilla.redhat.com/show_bug.cgi?id=1388295
Also add some notes in the README on how to run infrastructure
playbooks.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-10-26 12:51:08 -05:00
Andrew Schoen bebf412c92 infrastructure-playbooks: fix syntax errors in all playbooks
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
2016-10-25 16:56:58 -05:00
Sébastien Han f49bf2832c rolling_update: improve variables import
we now have pointer to default role so we don't miss any of the
variables defined.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-10-06 14:08:04 +02:00
Sébastien Han ac2cb9ac2c upgrade: add custom timeout options
This commit introduces the ability to configure delays and retries for
cluster health checks, for both monitors and OSDs.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-10-03 11:27:02 +02:00
Sébastien Han 21356c653f rolling updates: remove mon compact command
Users have reported this task to hang. Since this command is not
required to perform the upgrade, we remove it.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-09-13 10:09:07 +02:00
Rachana Patel ad5805f03e rolling_update.yml will not work if cluster name is not 'ceph'. Adding --cluster will solve this problem
Fixes issue #969

Signed-off-by: Rachana Patel <rachana83.patel@gmail.com>
2016-09-07 15:38:58 -04:00
Sébastien Han fde819d1a8 create a directory for infrastructure playbooks
Since we have a couple of infrastructure related playbooks
(additionnally to the roles we are using to deploy Ceph), it makes sense
to have them located in a separate directory.

Signed-off-by: Sébastien Han <seb@redhat.com>
2016-08-17 11:53:34 +02:00