We have been using redundant definitions so that users can override
order but this is no longer required in recent upgrades.
Change-Id: Id60a655fcdbd05abd34165a5430ab1683bf7cfd5
In order to make easy orchestration on all OpenStack db-sync, add this
tag so people can use this tag in composition layer.
A use case it to set some orchestration to make sure MySQL Galera is
ready before running any Exec with this tag.
Change-Id: I2df5ae03c13b7bde00621a8b208e7204b4747713
Closes-Bug: #1755102
When you execute nova-manage commands, oslo logs to the following
location (file name is dynamically created based on command name).
/var/log/nova/nova-manage.log
Because puppet-nova is executing these commands as root,
nova-manage.log is owned by root, preventing the 'nova-manage
db archive_deleted_rows' entry in nova's crontab from executing.
Permission denied: '/var/log/nova/nova-manage.log'
This log file is also an outlier, as all other log files in
/var/log/nova/ are owned by nova:nova.
Similar issues are possible for other nova logs, if for example
a nova services is initially started manually as root, so the
ownership of all nova logs is corrected before configuring nova.
Co-Authored-By: Oliver Walsh <owalsh@redhat.com>
Co-Authored-By: Diana Clarke <diana.joan.clarke@gmail.com>
Co-Authored-By: Maciej Kucia <maciej@kucia.net>
Closes-Bug: #1671681
Change-Id: I0ca0110cbf9139c79074cf603dcab9135f96e765
This patch changes the default worker count from ::processorcount to the
new ::os_workers fact. ::os_workers is based on the number of processors
(currently cpu/4) but is capped at a maximum of 8 worker processors.
This is a much more reasonable default in general and prevents excessive
resource consumption on systems with a large number of CPUs.
Change-Id: Ieb69b5a6a50d37b0a49a8c9e57275673d60b5f10
Rather than use simple_cell_setup which expects that there are already
existing computes, this change uses map_cell0 & create_cell to setup
cell_v2. Once the computes are configured, the cell_v2 discover_hosts
should be used to finalized the installation.
In addition, the db syncs need to be reordered as the api db sync
should run before the the cell_v2 setup. The main db sync should run
after.
map_cell0/simple_cell_setup now uses main nova DB connection instead
of the api DB connection.
Change-Id: I591b451197dc3bd0783978f5e3d2b1c830afe54e
Closes-Bug: #1656276
Related-Bug: #1656673
Co-Authored-By: Alex Schultz <aschultz@redhat.com>
As Openstack projects continue to have longer database migration
chains, the Puppet default timeout of 300 seconds for an execution
is becoming too short a duration on some hardware, leading to timeouts.
As projects continue to add more migration scripts without "pruning"
the base, timeouts will continue to become more frequent unless
this time can be expanded.
The proposed change adds a configuration option db_sync_timeout
which maintains the Puppet default of 300 but allows TripleO to
change the timeout to something higher.
Change-Id: I6b30a4d9e3ca25d9a473e4eb614a8769fa4567e7
Add parameter to nova::db::sync class to allow end
users to add command line parameters to the db sync command.
Change-Id: I89dea475495d5903d953ff4492ad2e0f12a293c1
Partial-bug: #1472740
Switch Nova to use $::os_service_default
Change logging.pp, db.pp and tests.
Change-Id: I928a93534c6d27c020b7afb5b7dda32c379e9d62
Related-bug: #1515273
In order to standardize the way dbsync are run across our modules,
we create a new class nova::db::sync.
This class will be included if sync_db is enabled.
By making this transition the nova::db::sync class
can be returned by the ENC.
A use case would be in an highly available environment, with 3 galera
nodes, include nova::api on every node with sync_db set to false
and have the ENC return nova::db::sync just for one node.
Change-Id: I795178f83fe5b997a2574fe17263643688efc48c
Takashi Kajinami