Some services configuration are purged at step 3 and re-applied at step 3: lost of time
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tripleo |
Fix Released
|
High
|
Steven Hardy |
Bug Description
Some services like Nova, Neutron, Octavia have purge_config enabled by default, which means that all parameters not managed in a specific step will be removed:
Notice: /Stage[
It takes time to remove them and makes deployments longer to finish.
It was introduced by https:/
A solution could be to purge configs only when upgrading TripleO between major versions, and don't purge config during initial deployment or updates (loss of time).
For that, we would require https:/
Changed in tripleo: | |
milestone: | none → pike-2 |
Changed in tripleo: | |
assignee: | nobody → Steven Hardy (shardy) |
status: | Triaged → In Progress |
Additionally the purge is possibly causing service disruptions between step 3 and 4 on stack updates because configurations are being removed in step 3 (with service restarts) and readded (with service restarts)