3 subject="""comment 5"""
4 date="2019-11-13T21:42:58Z"
6 finishCommandActions is reaching the retry case, and STM deadlocks there.
7 The WorkerPool is getting into a state where allIdle is False, and is not
8 leaving it, perhaps due to an earlier STM deadlock. (There seem to be two
11 Also, I notice with --json-error-messages:
13 {"command":"get","note":"from origin...\nchecksum...","success":false,"key":"SHA256E-s524288--07854d2fef297a06ba81685e660c332de36d5d18d546927d30daad6d7fda1541","error-messages":["git-annex: thread blocked indefinitely in an STM transaction"],"file":"1"}
15 So the thread that actually gets to run on the key is somehow reaching a
18 Which made me wonder if that thread deadlocks on enteringStage.
19 And it seems so. If Command.Get is changed to use commandStages
20 rather than transferStages, the test case succeeds.
22 Like finishCommandActions, enteringStage has a STM retry if it needs to
23 wait for something to happen to the WorkerPool. So again it looks like
24 the WorkerPool is getting screwed up.