Details
-
Bug
-
Resolution: Fixed
-
Major
-
None
-
None
-
None
-
None
-
0
-
SDK04
Description
When we integrated the transactions into the client, the logging probably didn't quite get refactored as it needed to be. Ideally, we have to only create a logger once, and everyone uses that log level, etc... Lets revisit this and make the changes. The transactions code has a few loggers (to separate out the txn stuff from the cleanup threads). Perhaps that can stay, perhaps we need to do this differently? Also – while looking, review the info in the logs, consider what could/should be there or not there, etc...
For instance, the transactions log a lot at info level at shutdown, perhaps that is debug or trace. And when we create a transactions object (automagically when we create the cluster), it logs its creation at info with the version info, etc... Probably the cluster should do that at info level, but transactions don't really need to at all, or at least at trace/debug...
Attachments
Issue Links
- blocks
-
PYCBC-1440 Update logging to get latest from cxx client
- Closed