https://pantsbuild.org/ logo
#general
Title
# general
i

important-librarian-62877

06/06/2020, 12:17 AM
Curious if anyone else has seen this?
Copy code
00:13:05:335 [DEBUG] pants.engine.internals.scheduler:pid=15743: computed 1 nodes in 1.049953 seconds. there are 12121 total nodes.
00:13:05:357 [ERROR] Build graph construction failed: ExecutionError 1 Exception encountered:
  Exception: Snapshot failed: Throw(Error { kind: Io(Os { code: 28, kind: Other, message: "No space left on device" }), paths: [] }, "<pants native internals>")
I did check and I have plenty of space on my drive.
h

hundreds-father-404

06/06/2020, 12:19 AM
I think we have seen this but not in a long time. @hundreds-breakfast-49010 do you remember seeing this?
What command were you running?
h

hundreds-breakfast-49010

06/06/2020, 12:20 AM
I somewhat regularly run out of space in my /tmp dir and get errors similar to this when running large pants jobs
i.e. the thing we have https://pants.readme.io/docs/troubleshooting to mitigate
👍 1
h

hundreds-father-404

06/06/2020, 12:21 AM
Ah, so the “Using an alternative tmpdir” suggestion, right?
h

hundreds-breakfast-49010

06/06/2020, 12:21 AM
yes
I'm not sure if that's what @important-librarian-62877 is running into but that error message is consistent with it
I think that this is more or less expected - when we're running tests on, say, the entirety of the pants codebase, there's legitimately enough parallelism that it's easy for there to be enough separate copies of the test code that pants needs more than 4G of scratch space
i

important-librarian-62877

06/06/2020, 12:26 AM
I was running our test-all script which underneath is just running
pants test ::
, so this seems to align. I'll give your suggestion a shot.
The above suggestion did not fix the problem. I happened to roll back to 1.26 from 1.28 and this problem does not occur.
3 Views