Comment 27 for bug 1589433
Revision history for this message
| Alex Schultz (alex-schultz) wrote Re: Deployment of env with mongodb and ceilometer is failed | #27 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
1b1ed1a
(Get the code!)
So there are two options.
1) flush memcache during the deployment after endpoint creation to make sure we don't get a stale catalog
2) add cleanup logic to upload_cirros with extra retries & retry delays to work around the stale catalog issue.
ETA for initial code for both items is EOD US 6/9
There are possible risks with both items:
The risks for flushing memcache is that there may be performance issues related to the complete flushing of memcache when scaling the cluster.
The risks for the additional timeouts is that the upload_cirros.rb script may be relying on the current implementation and by changing the flow of the script to do cleanups we might be adding additional failures in other scenarios.