| commit | 8d55eb284014a82dd49afceb34b35e15afdd8825 | |
| author | Lucian Wischik <ljw@fb.com> | |
| Wed, 14 Sep 2022 17:32:41 +0000 (14 10:32 -0700) | ||
| committer | Facebook GitHub Bot <facebook-github-bot@users.noreply.github.com> | |
| Wed, 14 Sep 2022 17:32:41 +0000 (14 10:32 -0700) | ||
| tree | 6c5850809c2d82cddda85d28cc00662a0ddfb430 | treesnapshot (tar.gz zip) |
| parent | 46063d3cc77ae94d6615f3bbcbb4b94f20579811 | commitdiff |
re-enable legacy hh_fanout tests
Summary:
Every diff I write, I see that legacy hh_fanout tests have been disabled. They were disabled in D31408593 (Oct 2021) due to flakiness. But I wonder just how flakey they are? It'd be nice to re-enable them and no longer see the warning. Also, I've been fixing some other long-standing flakey tests recently so I now have a bit more experience in investigating flakiness.
I've added two tags, "heavyweight" and "slow". This is so that TPX doesn't schedule so many other tests concurrently with these, and gives them more time.
I've deleted a load of test-scenarios of hh_fanout that we don't actually care about. Hh_fanout has only one single user, zoncolan. They solely use it in the simplest mode, which is start it up and query for the fanout of a file. Nothing more. So I deleted all scenarios that involve calculating errors, or doing incremental changes and further fanouts.
I've cc'd on this diff the oncoming oncalls thomasjiang, madgen, losvald, so that if they get tasks associated with this flakey test then they'll know to blame me and send the task my way right away.
Reviewed By: shayne-fletcher
Differential Revision: D39419387
fbshipit-source-id: be665a07340427f23a6cf6d3914826597de546be
Summary:
Every diff I write, I see that legacy hh_fanout tests have been disabled. They were disabled in D31408593 (Oct 2021) due to flakiness. But I wonder just how flakey they are? It'd be nice to re-enable them and no longer see the warning. Also, I've been fixing some other long-standing flakey tests recently so I now have a bit more experience in investigating flakiness.
I've added two tags, "heavyweight" and "slow". This is so that TPX doesn't schedule so many other tests concurrently with these, and gives them more time.
I've deleted a load of test-scenarios of hh_fanout that we don't actually care about. Hh_fanout has only one single user, zoncolan. They solely use it in the simplest mode, which is start it up and query for the fanout of a file. Nothing more. So I deleted all scenarios that involve calculating errors, or doing incremental changes and further fanouts.
I've cc'd on this diff the oncoming oncalls thomasjiang, madgen, losvald, so that if they get tasks associated with this flakey test then they'll know to blame me and send the task my way right away.
Reviewed By: shayne-fletcher
Differential Revision: D39419387
fbshipit-source-id: be665a07340427f23a6cf6d3914826597de546be
12 files changed: