Hi
For the same job, I may got 'complete'(for most case) and 'incomplete' in lava server dashboard, the logs are almost the same and both show that the job finished correctly and the result is pass.
My question is: how is the server define the job as 'incomplete'?
When I check the lava server log, I got the same error log for ALL 'complete' and 'incomplete' jobs(most of them are actually mark as complete in the dashboard result):
2020-10-10 09:40:58,386 DEBUG - rv1126-evb 2020-10-10 09:40:58,395 DEBUG -> rv1126-evb-01 (Idle, Unknown) 2020-10-10 09:40:58,395 DEBUG |--> [140442] scheduling 2020-10-10 09:40:58,486 INFO [140442] START => lava-slave247 (rv1126-evb-01) 2020-10-10 09:40:58,517 INFO [140442] lava-slave247 => START_OK 2020-10-10 09:41:09,539 DEBUG lava-logs => PING(20) 2020-10-10 09:41:13,537 INFO [140442] lava-slave247 => END (lava-run crashed, mark job as INCOMPLETE) 2020-10-10 09:41:13,759 ERROR [140442] Unable to dump 'description.yaml' 2020-10-10 09:41:13,759 ERROR [140442] Compressed data ended before the end-of-stream marker was reached Traceback (most recent call last): File "/usr/lib/python3/dist-packages/lava_server/management/commands/lava-master.py", line 312, in _handle_end description = lzma.decompress(compressed_description) File "/usr/lib/python3.5/lzma.py", line 340, in decompress raise LZMAError("Compressed data ended before the " _lzma.LZMAError: Compressed data ended before the end-of-stream marker was reached 2020-10-10 09:41:13,763 DEBUG lava-slave241 => PING(20)
Thanks,
- Kever
Hello,
Le sam. 10 oct. 2020 à 05:26, Kever Yang kever.yang@rock-chips.com a écrit :
Hi
For the same job, I may got 'complete'(for most case) and
'incomplete' in lava server dashboard, the logs are almost the same and both show that the job finished correctly and the result is pass.
My question is: how is the server define the job as 'incomplete'?
A job is marked as "complete" if LAVA was able to run every actions that are defined in the job definition (even deploy, boot and test actions). The result of the specific tests that are ran does not matter. If LAVA is able to run the tests, then the job is complete. Users should look at individual test results to determine if the tests was successful or not.
When I check the lava server log, I got the same error log for ALL
'complete' and 'incomplete' jobs(most of them are actually mark as complete in the dashboard result):
2020-10-10 09:40:58,386 DEBUG - rv1126-evb 2020-10-10 09:40:58,395 DEBUG -> rv1126-evb-01 (Idle, Unknown) 2020-10-10 09:40:58,395 DEBUG |--> [140442] scheduling 2020-10-10 09:40:58,486 INFO [140442] START => lava-slave247 (rv1126-evb-01) 2020-10-10 09:40:58,517 INFO [140442] lava-slave247 => START_OK 2020-10-10 09:41:09,539 DEBUG lava-logs => PING(20) 2020-10-10 09:41:13,537 INFO [140442] lava-slave247 => END (lava-run crashed, mark job as INCOMPLETE) 2020-10-10 09:41:13,759 ERROR [140442] Unable to dump 'description.yaml' 2020-10-10 09:41:13,759 ERROR [140442] Compressed data ended before the end-of-stream marker was reached Traceback (most recent call last): File "/usr/lib/python3/dist-packages/lava_server/management/commands/lava-master.py",
line 312, in _handle_end description = lzma.decompress(compressed_description) File "/usr/lib/python3.5/lzma.py", line 340, in decompress raise LZMAError("Compressed data ended before the " _lzma.LZMAError: Compressed data ended before the end-of-stream marker was reached 2020-10-10 09:41:13,763 DEBUG lava-slave241 => PING(20)
For some reeason, lava is unable to send "description.yaml" from the slave to the master. This should not be an issue in most cases but still strange.
Rgds
lava-users@lists.lavasoftware.org