[nailgun] [stats] ERROR update_action_log failed
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Fix Released
|
Medium
|
Artem Roma |
Bug Description
{"build_id": "2014-12-
1. Create environment (I tried with different simple configurations: 2-3 nodes, ubuntu, neutron, no additional componens).
2. Start deployment of environment.
3. ~10-15 minutes later such message appears:
2015-01-05 15:42:15 ERROR
[7f8c840ff740] (helpers) update_action_log failed: Instance '<Task at 0x45985d0>' has been deleted, or its row is otherwise not present.
4. Environment is deployed successfully.
Changed in fuel: | |
status: | New → Confirmed |
tags: | added: stats |
summary: |
- [nailgun] ERROR update_action_log failed + [nailgun] [stats] ERROR update_action_log failed |
Changed in fuel: | |
assignee: | Fuel Python Team (fuel-python) → Artem Roma (aroma-x) |
After inspection of logs from attached snapshot and many attempts to reproduce on local dev env I want to say that description of this bug is not precisely correct as the problem occurs when user try to delete empty cluster via UI.
Main reason of such issue is that in described case we end up operating inside of nailgun task manager code with task object which corresponding entry has been deleted from env. To be more precise: there is action_log updating routine in _call_silently method that exploits the instance's "cashe" attribute marked as deffered in orm terms (meaning request to db is performed when access to attribute established) after deletion of task on which the routine is dependent.
I will provide fix for the bug in a short time.