![Doug Wiegley](/assets/img/avatar_default.png)
In theory, two parallel threads that need backoff, with the same backoff timers, can synchronize and just repeatedly hose themselves. oslo.db already has a way to put some randomness into the retry logic, so let's try that out. Change-Id: I2090eb1ca5f3c09500069309d147880081b8213e
6 lines
197 B
YAML
6 lines
197 B
YAML
---
|
|
fixes:
|
|
- The oslo.db wrap_db_retry function now supports randomized time jitter
|
|
in its retry algorithm. Add support for that feature, if the installed
|
|
version of oslo.db supports it.
|