While troubleshooting some issues with our Redis connection we came across quite some cases that lead or could lead to stability issues.
The following changes are made:
Remove any existing socket files on startup
Add an automatic forced connection reconnect interval (enforces an expiration on existing redis sessions to ensure that any partially broken connections will be reconnected, e.g. master becoming read-only slave)
Increased default diagnostic connection check interval and make it configurable (the default is really resource intensive if there is an outage of the redis server(s))
Some logging and improvements
This PR is based on the updated Go structure from our PR https://github.com/salesforce/rmux/pull/26. In case you are not interested in the new Go structure we could rebase this PR on your master.
While troubleshooting some issues with our Redis connection we came across quite some cases that lead or could lead to stability issues.
The following changes are made:
This PR is based on the updated Go structure from our PR https://github.com/salesforce/rmux/pull/26. In case you are not interested in the new Go structure we could rebase this PR on your master.