Details
-
Bug
-
Status: Closed (View Workflow)
-
Blocker
-
Resolution: Fixed
-
2.4.0, 2.4.1, 2.4.2
-
None
Description
This is likely a regression in 2.4, as with 2.3 everything work as expected.
For a read-write split (at least), both maxctrl and maxadmin show negative number of connections to a service (the total count of connections seems correct and incremented in time).
Connections to a server are shown properly.
Output from maxctrl:
[root@cgdsqlmax1 assen.totin]# maxctrl list services
|
┌─────────────────────┬────────────────┬─────────────┬───────────────────┬──────────────────────────┐
|
│ Service │ Router │ Connections │ Total Connections │ Servers │
|
├─────────────────────┼────────────────┼─────────────┼───────────────────┼──────────────────────────┤
|
│ MaxAdmin-Service │ cli │ 0 │ 0 │ │
|
├─────────────────────┼────────────────┼─────────────┼───────────────────┼──────────────────────────┤
|
│ 172.16.69.4-Service │ readwritesplit │ -837 │ 151603 │ 172.16.69.2, 172.16.69.3 │
|
└─────────────────────┴────────────────┴─────────────┴───────────────────┴──────────────────────────┘
|
Same, but for servers - connections are shown properly:
[root@cgdsqlmax1 assen.totin]# maxctrl list servers
|
┌─────────────┬─────────────┬──────┬─────────────┬─────────────────┬──────────────┐
|
│ Server │ Address │ Port │ Connections │ State │ GTID │
|
├─────────────┼─────────────┼──────┼─────────────┼─────────────────┼──────────────┤
|
│ 172.16.69.2 │ 172.16.69.2 │ 3306 │ 4 │ Master, Running │ 0-1-32720771 │
|
├─────────────┼─────────────┼──────┼─────────────┼─────────────────┼──────────────┤
|
│ 172.16.69.3 │ 172.16.69.3 │ 3306 │ 4 │ Slave, Running │ 0-1-32720771 │
|
└─────────────┴─────────────┴──────┴─────────────┴─────────────────┴──────────────┘
|
This needs a fix as it breaks any attempt to monitor the load on a service in terms of concurrent connections.