Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Won't Do
-
2.4.7
-
None
Description
When executing a cluster sync, the passwords are not updated as is described in MXS-2948.
Original description:
Folks,
Setting up two MaxScale instances to be available in an active-passive with keepalived, I use to set up the first one with a basic configuration file and then, use dynamic commands to create the monitor, servers, the service and the listener.
After that, I copy the maxscale.cnf and the .secrets to the other server and, using the maxctrl cluster sync command, I sync all configurations and get the second instance up and running quickly.
root@max01: ~ # maxctrl --hosts 192.168.102.228:8989 cluster sync 192.168.102.227:8989 |
OK,OK,OK,OK,OK,OK
|
|
The case is that, I executed the above steps and got the password variable on the second host as ******, as you can see below, for the monitor definition:
root@max02: maxscale.cnf.d # cat cluster-monitor.cnf |
[cluster-monitor]
|
type=monitor
|
module=mariadbmon
|
user=maxmon
|
password=*****
|
[...snip...]
|
replication_user=mariadb
|
replication_password=*****
|
[...snip...]
|
replication_master_ssl=false |
assume_unique_hostnames=true |
enforce_simple_topology=false |
With that, I had an issue when MaxScale wasn't able to connect to backends to pull states:
root@max02: maxscale # maxctrl list servers |
┌─────────────┬─────────────────┬──────┬─────────────┬──────────────────┬──────┐
|
│ Server │ Address │ Port │ Connections │ State │ GTID │
|
├─────────────┼─────────────────┼──────┼─────────────┼──────────────────┼──────┤
|
│ server20 │ 192.168.102.230 │ 3306 │ 0 │ Auth Error, Down │ │ |
├─────────────┼─────────────────┼──────┼─────────────┼──────────────────┼──────┤
|
│ server21 │ 192.168.102.231 │ 3306 │ 0 │ Auth Error, Down │ │ |
└─────────────┴─────────────────┴──────┴─────────────┴──────────────────┴──────┘
|
|
Can you take a look? Thanks!
Attachments
Issue Links
- relates to
-
MXS-3560 MaxScale Cluster Configuration Sync
-
- Closed
-