| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Handlers normally only trigger at the end of the play, but in this case
we just set our node schedulable again resulting in it immediately
getting taken down again.
|
|
|
|
|
|
|
| |
Previously we were setting schedulability to the state defined in the inventory
without regard to whether or not it was manually made schedulable or
unschedulable. The right thing seems to be to record the state prior to upgrade
and set it back.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Prevents the network egress bug causing node restart to fail during 3.3
upgrade. (even though a separate fix is incoming for this)
Only catch is preventing the openshift_cli role, which requires docker,
from triggering a potential upgrade, which we still don't want at this
point. To avoid we use the same variable to protect docker installed
version as we use in pre.yml.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In a parallel step prior to real upgrade tasks, clear out all unused
Docker images on all hosts. This should be relatively safe to interrupt
as no real upgrade steps have taken place.
Once into actual upgrade, we again clear all images only this time with
force, and after stopping and removing all containers.
Both rmi commands use a new and hopefully less error prone command to do
the removal, this should avoid missed orphans as we were hitting before.
Added some logging around the current image count before and after this
step, most of them are only printed if we're crossing the 1.10 boundary
but one does not, just for additional information in your ansible log.
|
| |
|
| |
|
|
Refactored the 3.2 upgrade common files out to a path that does not
indicate they are strictly for 3.2.
3.3 upgrade then becomes a relatively small copy of the byo entry point,
all calling the same code as 3.2 upgrade.
Thus far there are no known 3.3 specific upgrade tasks. In future we
will likely want to allow hooks out to version specific pre/upgrade/post
tasks.
Also fixes a bug where the handlers were not restarting
nodes/openvswitch containers doing upgrades, due to a change in Ansible
2+.
|