Github user aledsage commented on the pull request:
https://github.com/apache/brooklyn-server/pull/150#issuecomment-222639688
Speaking to @bostko off-line, it seems that things are more complicated
when it comes to a windows script kicking off a reboot. The reboot might not
happen immediately.
The install script that triggers the reboot might return, and then we'll
start executing the subsequent scripts. But in the background, the windows VM
is doing stuff and is about to reboot. So it's a race as to which script we
will be executing when the reboot actually happens. If it's not the install
script, then the configuration option added in this PR will not be enough. The
provisioning of the windows blueprint might then fail at the post-intall step
(or uploading of runtime-files, or customize or launch, etc).
If it is a subsequent step that fails, we really would need to retry it.
But it might well not be idempotent!
It would be good to have a more holistic approach to fix this (and/or a
consistent approach applied to all the steps).
Perhaps we need a "reboot expected" configuration option. This would be
similar to `install.reboot.required`, but would not trigger the reboot. Instead
it would wait for something that looks like a reboot (i.e it going offline and
then back online), before continuing with the subsequent tasks.
@nakomis what are your thoughts?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
