Support for container stub drive reuse

[Zyqsempai]

Signed-off-by: bpopovschi [email protected]

Issue #243

Added possibility to reuse stub drives of just deleted containers:

• The Unmount() method was added to agent/drive_handler to support unmounting used stub drives.
• The Release() method was added to aggregate all steps we need to release used stub.
• Modified task Delete() method to use new Release() method.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[Zyqsempai]

@sipsma PTAL and maybe you will have an idea how to fix test, it looks like hostPath is missing, but I have no clue how to set it in that test.

[sipsma]

Apologies for the delayed review, I was on vacation the end of last week.

I took a first pass look and left some comments. I'm going to take another look after thinking about the corner cases surrounding what we should do if Release fails and will get back to on your question about the test at that time.

[sipsma]

Left a few more comments, for some reason the formatting error I commented on below are preventing the tests from running now (./drive_handler.go:112:15: Errorf format %s reads arg #1, but call has 0 args), so I'm not able to see the error you were asking about earlier. Once that's fixed I'll be able to take a look again though.

[sipsma]

Looking good! A few more comments but on a high-level everything is coming together. One more thing to keep in mind is that we'll want some integ-test coverage for this. Luckily, we already have an integ test, TestCreateTooManyContainers, that verifies the previous behavior of failing when you try to create a 3rd container after creating 2 containers in a 2-container VM and deleting one. So, you should be able to just update that test (and change its name) to just verify that now the behavior actually works. Let us know if you run into any trouble with updating that or have questions about the integ tests.

[Zyqsempai]

@sipsma PTAL

[sipsma]

LGTM! Please do a rebase on master as I think this is a a decent chunk of commits behind (but I'm not foreseeing any major conflicts), after that should be good to approve

[Zyqsempai]

@sipsma Done!

[sipsma]

Thanks again!

[kzys]

LGTM. Thanks!