cluster/wait-filter¶
Cluster UUID that created the alert
The Task cluster-wait-for-members
requires a filter condition to match
a Machine object against to determine if the Machine has either succeeeded
or failed the wait conditions.
By default the filter wait condition is set to:
Or(Runnable=Eq(false),WorkflowComplete=Eq(true))
For Clusters that utilize an OS installable Pipeline, the Machine objects will
be set to Runnable: false
when they reboot in to the OS Installer. in
those cases; the cluster/wait-filter
must be adjusted to not fail out
when the Runnable: false
condition is set (as expected).
For OS installable Pipelines the following filter MAY work. Future
DRP versions may have better Machine fields to supporting this, rather
than relying on overloading the Runnable
field.
And(Workflow=Eq(universal-runbook),WorkflowComplete=Eq(true),BootEnv=Eq(local))
Note
The above filter assumes that universal-runbook
is the final chained
workflow in the Pipeline. This must be adjusted to be the correct final
chained workflow.
Note
The above filter will NOT correctly catch a machine that failures
out due to Workflow/Task errors setting the machine correctly to
Runnable: false
. The Pipeline/Workflow chains must succeed
for the cluster to complete building.