Bug 1949943 - race in logical switch and port creation causes port creation to fail
Summary: race in logical switch and port creation causes port creation to fail
Keywords:
Status: CLOSED DEFERRED
Alias: None
Product: Red Hat OpenStack
Classification: Red Hat
Component: python-networking-ovn
Version: 16.1 (Train)
Hardware: Unspecified
OS: Unspecified
high
high
Target Milestone: beta
: 18.0
Assignee: Terry Wilson
QA Contact: Eran Kuris
URL:
Whiteboard:
Depends On:
Blocks: 1956941
TreeView+ depends on / blocked
 
Reported: 2021-04-15 12:57 UTC by ffernand
Modified: 2024-12-20 19:54 UTC (History)
16 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
: 1956941 (view as bug list)
Environment:
Last Closed: 2023-10-16 14:57:21 UTC
Target Upstream Version:
Embargoed:

Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Issue Tracker OSP-3046 0 None None None 2021-11-17 09:37:09 UTC

Description ffernand 2021-04-15 12:57:54 UTC 
62cf1b5ec006489db99e2b0ebfb55f57] Mechanism driver 'ovn' failed in create_network_postcommit: KeyError: UUID('ddb5c499-a147-46a1-8df8-d70677cdfd5d')
2021-04-14 20:02:13.119 54 ERROR neutron.plugins.ml2.managers Traceback (most
Comment 1 Daniel Alvarez Sanchez 2021-04-15 13:21:54 UTC 
Let me clarify some details because I believe that you hit the 'send' button way too early.


If we have a set of operations such as:

1) Create network
2) Create subnet
3) Create FIP
4) Create port
5) Associate FIP to port

It looks like under heavy load, the FIP association - step 5) - may fall into some Neutron worker that hasn't already seen that the port has been created in 4).

This is a legitimate race condition that we probably can fix via OneTime events on certain things such as the FIP association but we've seen this as well happening when one of the workers has a stale connection to NB database and hence not processing events and unable to commit transactions.

(Private) logs from what we saw yesterday could help address this last scenario.
Comment 2 Rodolfo Alonso 2021-04-15 13:28:41 UTC 

*** This bug has been marked as a duplicate of bug 1946262 ***
Note You need to log in before you can comment on or make changes to this bug.