Add cleanup mechanism to PyTorch org GHA runners

[malfet]

For example i-043e2307579751589 run out of disc space (as one can see from https://github.com/pytorch/tutorials/actions/runs/5245394136/jobs/9474787159 ) due to 72 corefiles in pytorch/test folder:

ec2-user@ip-10-0-2-80 test]$ ls core.* -lah
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:04 core.48879
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:04 core.48880
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:04 core.48913
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:04 core.48914
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:04 core.48947
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:04 core.48948
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:04 core.48982
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:04 core.49015
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:04 core.49016
-rw------- 1 ec2-user ec2-user 1.9G Jun 12 12:04 core.49084
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:05 core.49152
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:05 core.49185
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:05 core.49186
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:05 core.49253
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:05 core.49254
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:05 core.49287
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:05 core.49394
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:05 core.49427
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:05 core.49428
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:06 core.49461
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:06 core.49462
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:06 core.49496
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:06 core.49563
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:11 core.49598
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:11 core.49631
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:11 core.49665
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:11 core.49666
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:11 core.49734
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:11 core.49767
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:11 core.49768
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:12 core.49907
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:12 core.49908
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:12 core.49941
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:12 core.49942
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:12 core.49975
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:12 core.49976
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:12 core.50009
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:12 core.50010
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:12 core.50078
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:17 core.50111
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:17 core.50112
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:18 core.50213
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:18 core.50214
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:18 core.50247
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:23 core.50281
-rw------- 1 ec2-user ec2-user 1.9G Jun 12 12:23 core.50282
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:23 core.50315
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:23 core.50316
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:28 core.50387
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:28 core.50388
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:28 core.50422
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:33 core.50455
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:33 core.50456
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:33 core.50489
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:38 core.50523
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:38 core.50524
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:43 core.50744
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:44 core.50777
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:44 core.50778
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:44 core.50819
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:44 core.50861
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:44 core.50862
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:44 core.50896
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:44 core.50938
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:44 core.50979
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:44 core.50980
-rw------- 1 ec2-user ec2-user 1.7G Jun 12 12:44 core.51055
-rw------- 1 ec2-user ec2-user 1.8G Jun 12 12:44 core.51056
-rw------- 1 ec2-user ec2-user 769M Jun 12 12:47 core.51101
-rw------- 1 ec2-user ec2-user 886M Jun 12 12:47 core.51102
-rw------- 1 ec2-user ec2-user    0 Jun 12 12:47 core.51211
-rw------- 1 ec2-user ec2-user    0 Jun 12 12:47 core.51212

Would be nice to have a script that deletes core files (or cleans up repo folder after action is finished)

[huydhn]

IMO, having a cleanup script works in most situation, but there are some edge cases like the above example to keep in mind. The runner started to run out of space in https://github.com/pytorch/pytorch/actions/runs/5241340530/jobs/9463851052 in the middle of testing, and any cleanup script that is part of the workflow wouldn't even have the chance to run. Then the next job like https://github.com/pytorch/tutorials/actions/runs/5245394136/jobs/9474787159 wouldn't be able to checkout the action to run any cleanup because the runner has already ran out of space. So there would also need to be a ulimit -c to limit the size core dumps or a clean up cron job.