Project

General

Profile

Actions
Copy link

action #37312

closed

Workers were deleted and recreated

Added by mloviska almost 6 years ago. Updated about 5 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
mkittler
Category:
Feature requests
Target version:
Current Sprint
Start date:
2018-06-13
Due date:
% Done:

0%

Estimated time:

Description

I have noticed that we are randomly missing Assigned worker information in ppc jobs, although Logs & Assets are still accessible.

Expected results:

Missing data:

Actions
Copy link
 #1

Updated by coolo almost 6 years ago

something/someone deleted the worker from the database. The QA-Power8-4-kvm worker we have right now was created 2018-06-04 - aka after your jobs ran.

And deleting worker DB entries removes the reference in jobs.

But I wonder if someone really deleted this manually, e.g. worker id 1073 was created just yesterday, even though the systemd service
for it has logs for May 20.

Actions
Copy link
 #2

Updated by coolo almost 6 years ago

  • Subject changed from [tools] Missing assigned worker in result bar to [tools] Workers were deleted and recreated
  • Priority changed from Normal to Low

the workers weren't removed since then. Still no idea what triggered the deletion, but it's on the monitoring list

Actions
Copy link
 #3

Updated by coolo over 5 years ago

  • Subject changed from [tools] Workers were deleted and recreated to Workers were deleted and recreated
  • Category set to 168
  • Priority changed from Low to Normal
  • Target version set to Ready

It's still happening - notice the vastly different t_created

openqa=> select * from workers where host='openqaworker5' order by instance;
  id  |     host      | instance |      t_created      |      t_updated      | job_id  | upload_progress 
------+---------------+----------+---------------------+---------------------+---------+-----------------
  957 | openqaworker5 |        1 | 2018-03-20 13:11:06 | 2018-10-18 08:48:26 | 2187481 | 
  961 | openqaworker5 |        2 | 2018-03-20 13:15:53 | 2018-10-18 08:48:26 | 2187308 | 
  985 | openqaworker5 |        3 | 2018-03-20 13:26:13 | 2018-10-18 08:48:18 | 2187530 | 
 1202 | openqaworker5 |        4 | 2018-10-11 05:10:25 | 2018-10-18 08:48:24 | 2187185 | 
  984 | openqaworker5 |        5 | 2018-03-20 13:25:26 | 2018-10-18 08:48:25 | 2186488 | 
  660 | openqaworker5 |        6 | 2017-07-24 17:23:19 | 2018-10-18 08:48:23 | 2186498 | 
  355 | openqaworker5 |        7 | 2016-11-13 18:24:14 | 2018-10-18 08:48:20 | 2187209 | 
  646 | openqaworker5 |        8 | 2017-07-24 17:08:07 | 2018-10-18 08:48:22 | 2187229 | 
  357 | openqaworker5 |        9 | 2016-11-13 18:24:16 | 2018-10-18 08:48:24 | 2186671 | 
  960 | openqaworker5 |       10 | 2018-03-20 13:15:46 | 2018-10-18 08:48:23 | 2187430 | 
  918 | openqaworker5 |       11 | 2017-12-21 09:53:21 | 2018-10-18 08:48:21 | 2187184 | 
  847 | openqaworker5 |       12 | 2017-12-21 09:42:35 | 2018-10-18 08:48:26 | 2187554 | 
  958 | openqaworker5 |       13 | 2018-03-20 13:11:08 | 2018-10-18 08:48:19 | 2187520 | 
  895 | openqaworker5 |       14 | 2017-12-21 09:48:06 | 2018-10-18 08:48:21 | 2187453 | 
 1193 | openqaworker5 |       15 | 2018-10-11 05:10:08 | 2018-10-18 08:48:26 | 2187429 | 
  953 | openqaworker5 |       16 | 2018-03-20 13:07:18 | 2018-10-18 08:48:26 | 2187416 | 
  967 | openqaworker5 |       17 | 2018-03-20 13:16:06 | 2018-10-18 08:48:15 | 2187657 | 
  955 | openqaworker5 |       18 | 2018-03-20 13:08:10 | 2018-10-18 08:48:19 | 2187572 | 
 1203 | openqaworker5 |       19 | 2018-10-11 05:10:27 | 2018-10-18 08:48:19 | 2186499 | 
  669 | openqaworker5 |       20 | 2017-07-26 12:47:24 | 2018-10-18 08:48:21 | 2187315 |
Actions
Copy link
 #4

Updated by mkittler about 5 years ago

I've just been looking at the code and it seems we don't provide an API route for deleting workers. It is hard to tell for sure but it also doesn't look like that the openQA code ever deletes workers (at least not via DBIx delete).

Maybe some SQL cascade delete does that?

Actions
Copy link
 #5

Updated by coolo about 5 years ago

hmm, e.g. when deleting running jobs?

Actions
Copy link
 #6

Updated by mkittler about 5 years ago

  • Status changed from New to In Progress
  • Assignee set to mkittler
  • Target version changed from Ready to Current Sprint

Could be. I'll try to reproduce that with the test fixtures.

Actions
Copy link
 #7

Updated by mkittler about 5 years ago

Actions
Copy link
 #8

Updated by coolo about 5 years ago

Be careful - let's say it's one problem :)

But I'm confident

Actions
Copy link
 #9

Updated by mkittler about 5 years ago

Right now I can only think of one more relation. But I've just did a quick test with fixtures and there's no deletion happening. If you like we can add this test nevertheless: https://github.com/Martchus/openQA/pull/new/test_previous_job_deletion

Actions
Copy link
 #10

Updated by mkittler about 5 years ago

  • Status changed from In Progress to Resolved
Actions
Copy link

Also available in: Atom PDF