-
Type: Task
-
Resolution: Done
-
Priority: Major - P3
-
None
-
Affects Version/s: 2.7.0
-
Component/s: Server Selection
-
None
-
(copied to CRM)
The driver's topology flaps between known and unknown states, at approximate intervals of 30seconds, then 7 seconds alternating. The 7 seconds corresponds to the configured parameter "server_selection_timeout: 7". Coincidentally the 7 seconds is also used in the test harness for rapid failover testing, and does not appear to have been updated to reflect the SDAM revamp in 2.7.0.
I suspect this might be in part addressed by the work on RUBY-1434. Though I wanted to bring this extra information to you attention, in addition to make the suggestion that the bounds check for the server_selection_timeout parameter be increased from 3 to something approaching 10-30 seconds.
- is related to
-
RUBY-1650 Rewrite connection handshake flow to not recurse
- Closed
-
RUBY-1715 Log ismaster retries on monitoring connections
- Closed
-
SERVER-40583 Provide reason for why connections are closed
- Closed
-
RUBY-1434 Document low server selection timeout in development mode
- Closed
- related to
-
SERVER-39941 Report TLS handshake errors to client
- Backlog