You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If a node is placed just in the reader_hostgroup in a cluster configured with writer_is_also_reader=2, the node will be permanently dropped from the 'runtime_mysql_servers' table.
Force a server to be placed only in the reader_hostgroup, with: SET GLOBAL read_only=ON;. The server has now completely disappear from the runtime_mysql_servers table, leaving it like this:
mysql: [Warning] Using a password on the command line interface can be insecure.
+-----------+-------------+----------+---------+---------+
| hostgroup | srv_host | srv_port | status | Queries |
+-----------+-------------+----------+---------+---------+
| 20 | 172.18.1.11 | 3306 | SHUNNED | 0 |
| 20 | 172.18.1.13 | 3306 | ONLINE | 0 |
| 30 | 172.18.1.11 | 3306 | ONLINE | 0 |
| 10 | 172.18.1.11 | 3306 | ONLINE | 0 |
+-----------+-------------+----------+---------+---------+
Now, when trying to recover the sever, reverting the read_only config, the server never gets back, after setting: SET GLOBAL read_only=OFF;, the runtime_mysql_servers table keeps being:
mysql: [Warning] Using a password on the command line interface can be insecure.
+-----------+-------------+----------+---------+---------+
| hostgroup | srv_host | srv_port | status | Queries |
+-----------+-------------+----------+---------+---------+
| 20 | 172.18.1.11 | 3306 | SHUNNED | 0 |
| 20 | 172.18.1.13 | 3306 | ONLINE | 0 |
| 30 | 172.18.1.11 | 3306 | ONLINE | 0 |
| 10 | 172.18.1.11 | 3306 | ONLINE | 0 |
+-----------+-------------+----------+---------+---------+
The full ProxySQL error log (default location: /var/lib/proxysql/proxysql.log)
The attached tar contains:
The full ProxySQL error log.
The logs from the individual mysql containers during the testing.
The cluster configuration.
The final cluster status after resetting the read only flag.
If a node is placed just in the reader_hostgroup in a cluster configured with
writer_is_also_reader=2
, the node will be permanently dropped from the 'runtime_mysql_servers' table.v2.0.16
v2.1.0
Create a Galera cluster with the following config:
Initial
runtime_mysql_servers
looks like this:Force a server to be placed only in the
reader_hostgroup
, with:SET GLOBAL read_only=ON;
. The server has now completely disappear from theruntime_mysql_servers
table, leaving it like this:Now, when trying to recover the sever, reverting the read_only config, the server never gets back, after setting:
SET GLOBAL read_only=OFF;
, theruntime_mysql_servers
table keeps being:/var/lib/proxysql/proxysql.log
)The attached tar contains:
reader_gone_logs.tar.gz
The text was updated successfully, but these errors were encountered: