Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
With new pacemaker architecture, Puppet handles restarts of most of the
services. There are several still managed by pacemaker which need
special restart handling utilizing pacemaker and its resource
agents.
The counterpart in puppet-tripleo requests restarts for individual
pacemaker-managed services by writing out "restart flag" files, and the
pacemaker_resource_restart.sh script then performs the restarts.
Change-Id: Ia4e6a9f88181f1981993f046cf415dbbcdc9570e
Closes-Bug: #1614967
Closes-Bug: #1587015
Depends-On: I6369ab0c82dbf3c8f21043f8aa9ab810744ddc12
|
|
|
|
|
|
|
|
Adds a pre-requisite software deployment to the pacemaker scenario
upgrade which, before the openstack services are upgraded,
upgrades the ceph-mon daemon from Hammer to Jewel.
Change-Id: I9855d80a6ae156b4a9e0409c3c927068b9db95a0
|
|
|
|
|
|
We have to recreate the /var/lib/mysql directory on all controller node,
not just the boostrap node for the cluster to be able to restart.
Adding a warning on the fact that those script are local and know
nothing about the good upgrade state of the other nodes.
Closes-Bug: 1612642
Change-Id: I48e2812d7df80bbf2db53a8b71dc434d4209a160
|
|
There is a typo in the code, making this test always successful.
Closes-Bug: 1614437
Change-Id: Ia6b0b156294de9fcb8f66fc46aa8801555775a56
|
|
scheduler_host_manager doesn't take
nova.scheduler.host_manager.HostManager as a value anymore. This fix it
before restarting the service.
Change-Id: Ia9adcfd5a898f0c712b4a37ae33db88a44630f0d
Closes-Bug: 1615035
|
|
|
|
|
|
These are not needed any more as they were specific to
mitaka upgrades.
Change-Id: I0d421b942e620403f88374e1c82105747d8d84c9
|
|
The nova api db need to be synchronized as well.
Change-Id: I2628b24ff1153c84cbf388455666ae42570cb10f
Closes-Bug: 1615042
|
|
The MySqlMajorUpgrade parameter has validation on it allowing only
values yes/no/auto, however in the script we checked for '0' instead of
'no', which means the only effective values were yes/auto. This is now
fixed to allow switching the migration off.
Change-Id: I5d64734894c6bfd9003ad643f3747e34e62465cc
Closes-Bug: #1616429
|
|
When upgrading from mariadb X.Y.Z to mariadb X.Y.Ž (X.Y part stays the
same), the dump/restore of mariadb shouldn't be necessary. Therefore we
now only check for up to the first 2 fields of the version string when
determining if we should trigger the dump/restore operation.
Closes-Bug: #1615721
Change-Id: Ib7af8bfb121f5c83184d51b3c6dc657108c25973
|
|
In Newton, Aodh will be using its own mysql DB rather than
using ceilometer's mongo instance. This means we need to
migrate any existing alarm and alrm history data from
ceilometer DB to aodh mysqlDB. Upstream aodh provides us
with a aodh-data-migration utility. We need to invoke this
during the mitaka->newton upgrade procedure so data is
migrated as expected and aodh mysql backend takes over.
Closes-bug: #1611794
Change-Id: I17888b57ecf98cd83e92af2f9cdbead066b03aa3
|
|
Given the new HA architecture with less pacemaker managed resources, we
need to update this script to reflect those changes. Without these
changes, stack-updates using the exact same templates will fail since
this script is always executed on update.
Change-Id: I2ce1681d19d4a24a7561e3dd9c5efdae40d030b7
Closes-Bug: #1612667
|
|
|
|
|
|
When the overcloud is upgraded we do a yum update of the packages.
This step might introduce a newer galera version. In such a situation
we need to dump the db and restore it. The high-level workflow should
be the following:
1) During the main upgrade step, before shutting down the cluster
we need to dump the db
2) We upgrade the packages
3) We briefly start mysql on a single node while making sure that
/root/.my.cnf is briefly moved out of the way (because it contains
a password) and import the data. After the import we shutdown this
mysql instance
4) We let the cluster start up normally
The above steps will take place in the following scenarios.
Given a locally installed mariadb version X.Y.Z and release R,
we will dump and restore the DB under the following conditions:
A) MySqlMajorUpgrade template parameter is set to 'auto' and
the upgraded package differs in X, Y *or* Z. We basically don't
dump automatically if the release field changes.
B) MySqlMajorUpgrade template parameter is set to 'yes'
When MySqlMajorUpgrade is set to 'no', no dumping will be performed.
Note that this will give a non functional upgrade if a major mariadb
upgrade is taking place.
Partial-Bug: #1587449
Co-Author: Damien Ciabrin <dciabrin@redhat.com>
Co-Author: Mike Bayer <mbayer@redhat.com>
Depends-On: I8cb4cb3193e6b823aad48ad7dbbbb227364d2a58
Depends-On: I38dcacfabc44539aab1f7da85168fe44a1b43a51
Change-Id: I374628547aed091129d0deaa29764bfc998d76ea
|
|
Since the Liberty release, the number of services managed by pacemaker
on HA Overcloud has increased. This has an impact on
major_upgrade_controller_pacemaker_1.sh, where cluster sync timeout
value tuned for older releases is now becoming too low.
Raise the cluster sync timeout value to a sensible limit to
give pacemaker enough time to stop the cluster during major upgrade.
Change-Id: I821d354ba30ce39134982ba12a82c429faa3ce62
Closes-Bug: #1597506
|
|
It is best if we disable stonith if a cluster has it configured and on,
before we call "pcs cluster stop --all", because should a service fail
to stop for whatever reason, pacemaker will fence the node where it
happened. This is something that we unlikely want during an upgrade as
it will make things worse.
Once the cluster is stopped we can reenable stonith (if it was enabled
to start with) in the CIB while the cluster is shut down.
Closes-Bug: #1596065
Change-Id: I38dcacfabc44539aab1f7da85168fe44a1b43a51
|
|
Restarting services after Puppet is vital to ensure that config changes
go applied. However, it can be sometimes desirable to prevent these
restarts to avoid downtime, if the operator is sure that no config
changes need applying. This can be a case e.g. when scaling compute
nodes. Passing the puppet-pacemaker-no-restart.yaml environment file *in
addition* to puppet-pacemaker.yaml should allow this.
This is a stop gap solution before we have proper communication between
Puppet and Pacemaker to allow selective restarts.
Change-Id: I9c3c5c10ed6ecd5489a59d7e320c3c69af9e19f4
|
|
|
|
|
|
If "pcs cluster stop --all" is executed on a controller that
happens to have a VIP on the internal network, pcs may use the
VIP as the source address for communication with another cluster
node. When pacemaker is stopped this VIP goes away, and pcs never
receives a response from the other node. This causes pcs to hang
indefinitely; eventually the upgrade times out and fails.
Disabling the VIPs before stopping the cluster avoids this
situation.
Change-Id: I6bc59120211af28456018640033ce3763c373bbb
Closes-Bug: 1577570
|
|
|
|
|
|
Previously ceilometer-notification, aodh-listener and sahara-engine
didn't have constraints that would anchor them under openstack-core
dummy resource. Such constraints are added now. (sahara-engine starting
after sahara-api, aodh-listener after aodh-evaluator, and
ceilometer-notification after openstack-core.) Openstack-core ->
heat-api constraint has been removed because heat-api depends on
ceilometer-notification, so there's a transitive dependency on
openstack-core already.
Change-Id: Ided7321ebbf2c3556726343b4bb466fd8759b43a
Closes-Bug: #1569444
|
|
Previously we tried to use UpdateIdentifier for two different things:
tell whether to perform package update, and also to tell whether the
top-level stack is being created or updated (which was incorrect and
resulted in bug 1567384, and an attempt to work around that bug resulted
in bug 1567385).
We cannot use Heat's "action" conditionals in some cases, because they
refer to the direct parent stack, which can yield undesirable results
when introducing new nested stacks or temporarily no-opping something
and then adding it back (in both these cases, "action" would be
considered "CREATE", even though the top-level stack is in "UPDATE").
So tripleoclient passes a new parameter StackAction to tell whether the
top-level stack is being created or updated, and we make use of
that. (It seems there's no better way of getting this info from within
the nested Heat stacks.)
Change-Id: Ie14ddbff15e7ed21aaa3fcdacf36e0040f912382
Depends-On: I9dc3b4cd8a6a71df34d8babf0e4c6505041f5311
Closes-Bug: #1567384
Related-Bug: #1567385
|
|
The update and upgrade shell scripts were still referencing the
old openstack-keystone service which got removed with
Ie26908ac9bfc0b84b6b65ae3bda711236b03d9d4
Also removes kilo and liberty specific workarounds and config changes.
Change-Id: Icc80904908ee3558930d4639a21812f14b2fd12e
|
|
Removes the old noop nested stack template for extraconfig
tasks and instead uses OS::Heat::None. This should avoid a few
extra resource checks on create and update.
Change-Id: I5a42fc78ece2553e86385236e214aa1e3c91cd85
|
|
The change at https://review.openstack.org/#/c/302352/ should stop
the if up/down scripts from making changes to resolv.conf as
discussed in that review and the related bug below. However during
upgrades, as we are moving from a version of the ifcfg-vlanXX files
that don't have the PEERDNS=no added by /#/c/302352 the if up
script will restore the /etc/resolv.conf.save to /etc/resolv.conf
and overwrite it. This removes the .save file during the upgrade
init command which gets delivered to all nodes as the first stage
of a major upgrade.
Change-Id: I91dd139f43be4912c20d8661691bee2b662964d4
Related-Bug: 1567004
|
|
While having extra customizations inside a TripleO deployed
Pacemaker environment, say you have instance HA with
pacemaker_remoted or you need to configure an external arbitrator
for something, then the status of the resources for remote nodes
is "Stopped".
This leads to failures while, for example, scaling up.
This fixes the way status is checked, filtering just local nodes.
Co-Authored-By: Giulio Fidente <gfidente@redhat.com>
Change-Id: I8dc25f5d7031c265858afd5a266fda5315ae37a0
|
|
If a certificate expires, the user will need to update it. However,
because we only restart services at the end of a stack-update the
new certificate doesn't take effect until after puppet has run.
This is a problem because puppet makes OpenStack calls, which will
fail if the certificate is expired. In that case we never get to
the service restart so the stack is wedged until the user manually
restart haproxy.
This patch addresses the problem by reloading haproxy before puppet
runs. This is done in a pre-puppet script for pacemaker after pacemaker
is maintenance mode because we need to make sure it happens after all of
the certs have been installed on the controllers, but before puppet
runs.
For non-pacemaker, haproxy is simply reloaded.
Change-Id: Id5ed05b3a20d06af8ae7a3d6f859b03399b0d77d
|
|
We'd like to let the post puppet pacemaker controller services
restart to happen for the convergence step so set the
UpdateIdentifier. However also set the PackageUpdate to noop so the
yum_update.sh doesn't happen.
Since a full haproxy restart is expected, we no longer need the
systemctl reload added at Iae3bad745ecdf952a7a0314fe1375d07eb47c454
so remove that too.
Some more context at
https://bugzilla.redhat.com/show_bug.cgi?id=1321036
Co-Authored-By: marios <marios@redhat.com>
Change-Id: I31c2d97d68c97b435f63863fae2c89f18f99681d
|
|
|
|
|
|
As discussed in the related bug below, after upgrading your
environment to latest liberty the haproxy config isn't picked
up. This adds a systemctl reload haproxy in the pacemaker
resource restart we run as part of the post-puppet-pacemaker.
Related-Bug: 1561012
Change-Id: Iae3bad745ecdf952a7a0314fe1375d07eb47c454
|
|
Ceilometer Alarm is deprecated in Liberty by Aodh.
This patch:
* manage Aodh Keystone resources
* deploy Aodh API under WSGI, Notifier, Listener and Evaluator
* manage new parameters to customize Aodh deployment
* uses ceilometer DB for the upgrade path
* pacemaker config
* Add migration logic to remove pcs resources
Depends-On: I5333faa72e52d2aa2a622ac2d4b60825aadc52b5
Depends-On: Ib6c9c4c35da3fb55e0ca8e2d5a58ebaf4204d792
Co-Authored-By: Emilien Macchi <emilien@redhat.com>
Change-Id: Ib47a22884afb032ebc1655e1a4a06bfe70249134
|
|
|
|
|
|
|
|
|
|
Yum update on cinder nodes should be quiet, as it is on controllers,
because results of these updates are sent to Heat. I mistakenly left
this out in the first patch because i used one of the standalone node
upgrade scripts as a copy/paste base for the cinder node upgrade script.
Change-Id: Id13190dc4d242317829c7994088183f52d21461d
|
|
|
|
The variables in the heredoc should be escaped because they should
evaluate only when the inner script runs, not when the outer "writer"
script runs.
Python-zaqarclient is installed for os-collect-config to work, as we do
on the other node types.
Swift-proxy is removed from list of services to stop/start, as
swift-proxy isn't supposed to run on the swift storage nodes.
Change-Id: I8426b859d11378ebdc3da94dcc090133dab0c628
|
|
During the controller upgrade in
major_upgrade_controller_pacemaker_1.sh we use systemctl to stop
all swift services and then start them again in _pacemaker_2.sh
In the case of stand-alone swift nodes the deployer may have
used the ControllerEnableSwiftStorage: false so that only the
swift-proxy service is left on controllers (wrt swift). The
systemctl_swift function used during upgrades is changed to factor
this in.
Change-Id: Ib22005123429f250324df389855d0dccd2343feb
|