-
Notifications
You must be signed in to change notification settings - Fork 3
Conversation
Signed-off-by: Ken Sipe <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
looks good as first automation of install/delete for MWT using KUTTL
apiVersion: kuttl.dev/v1beta1 | ||
kind: TestStep | ||
commands: | ||
- command: kubectl kudo install cassandra --skip-instance --namespace cassandra |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we should test from the local path in this case if MWT scripts are living next to operator
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@zmalik I don't know what you mean
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
if you are referring to using the local build of this operator in this repo... I would rather not... MWT should only be on released operators and the scripts should work against the released operator repo IMO
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would generally agree, although with a project that is heavily in development this might not always be possible. For the last MWT I used a dev-version from master.
I do hope that we have reached a state where this won't be necessary anymore though.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we usually participate with the next unreleased version in MWT to make sure there are no stability issues with the upcoming release under load tests.
that's a non-blocker issue as would require minimal effort to change it when running MWT. Moreover, now given the new dates we will participate with a released version so each time it will heavily dependent on operator release dates and MWT dates
….com> Signed-off-by: Ken Sipe <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
lgtm, especially for a start. I think using kuttl for certain aspects of MWT has a merit
Signed-off-by: Ken Sipe <[email protected]>
….com> Signed-off-by: Ken Sipe <[email protected]>
Signed-off-by: Ken Sipe <[email protected]>
after fixing kudo... might want to look at
|
Signed-off-by: Ken Sipe <[email protected]>
updated to cassandra 1.0.0 + added test configs |
…erator into ken/mwt � Conflicts: � benchmarks/README.md � benchmarks/mwt/README.md � benchmarks/mwt/setup/01-cassandra-install/02-install-cassandra.yaml
Signed-off-by: Andreas Neumann <[email protected]>
Switch to using KUTTL for MWT
I've tested against a Konvoy cluster with the following changes to the
setup/01-cassandra-install/02-install-cassandra.yaml
fileI would be inclined to restructure the folders and README if we can get agreement on this is the way we want to move forward. (I would likely remove or rename
mwt
folder and merge the 2 README files)I did discover some weakness in kuttl with this... namely around namespace management... the
delete
properties for a TestStep do not honor the namespace defined... they only delete what is in the generated test namespace. Which has worked well for tests but isn't great when the focus is setup and verify. Created kudobuilder/kuttl#124 to track.Signed-off-by: Ken Sipe [email protected]