Support for ping check to get healthy connections #9
Conversation
Codecov Report
@@ Coverage Diff @@
## master #9 +/- ##
==========================================
- Coverage 84.02% 78.83% -5.19%
==========================================
Files 3 3
Lines 169 189 +20
==========================================
+ Hits 142 149 +7
- Misses 27 40 +13
Continue to review full report at Codecov.
|
| c.ping() | ||
| except: | ||
| self.current_size -= 1 | ||
| if retry_count < 10: c = self.__get_conn(retry_count+1) |
There was a problem hiding this comment.
Assume retry exceeds 10, ping() still fails. Is there need to remove this invalid conn and recreate a conn, or throw an exception?
There was a problem hiding this comment.
When the ping() fails, the conn is discarded and another conn is fetched from the pool. The retry_count is used to to get and return the first healthy conn from the pool a couple of times. If the conn object fails on trying reconnect, maybe the object has some problem with its internal properties. But... if other instances have the same problem, maybe there is a network problem on going, and no conn will be returned anymore. This implementation just return a None conn on a network problem occurring, but we can make a custom exception if you prefer.
There was a problem hiding this comment.
Oh, ping() will reconnect, I forgot it 😅...
Nice pr 👍🎉👏, merged.
Closes #8