Issue metadata
Sign in to add a comment
|
test-level retries for VMTest |
||||||||||||||||||||||||
Issue descriptionRetrying all of VMTests stage if a single test failed is slow, as evidenced by for instance https://viceroy.corp.google.com/chromeos/build_details?build_config=betty-paladin&build_number=226 Instead, we can push simple retry logic into test_that (which is what runs VMTests, on builder). Gotta be a bit careful how we do this so that we don't end making things worse (ie slower) by over-retrying.
,
Jul 26 2017
I assume that it's not reasonable to somehow have some unified test running experience that lets run tests and handle retries (or their configuration) in a similar fashion for VMTest and HWTest?
,
Jul 26 2017
The problem there wasn't a failing test, but that the VM disappeared and couldn't be connected to anymore. What is needed here is a pool of VMs (to run tests in parallel on the server) and VM recovery.
,
Aug 10 2017
,
Aug 18 2017
|
|||||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||||
Comment 1 by akes...@chromium.org
, Jul 26 2017