6c610e23ce
Add `designate-manage service clean` cmd to allow clean stale services. Any services not provide heartbeat within double heartbeat_interval time needs to be consider as dead services. And should be removed ASAP. And in any case, even service is actually alive (maybe in a crazy rpc timeout case), it can alway update and create it's service status back. So we're fine to delete the service status in this case too. Closes-Bug: #2110262 Change-Id: I211d0a60aa19ab4c0ae0fbb808e2d9080ccbbedd Signed-off-by: ricolin <rlin@vexxhost.com>
14 lines
556 B
YAML
14 lines
556 B
YAML
---
|
|
fixes:
|
|
- |
|
|
New cmd `designate-manage service clean`.
|
|
Previously Designate service always stuck in `UP` even we
|
|
stop provide any heartbeat for long while.
|
|
And no method to clean services status.
|
|
For services that run on containers (like in K8s),
|
|
The list of `UP` services statuses just piling up.
|
|
Add new cmd `designate-manage service clean` to delect and
|
|
clean any service that fail to provide heartbeat within
|
|
double heartbeat interval time.
|
|
`CONF.heartbeat_emitter.heartbeat_interval` default to 10 seconds.
|