reload startup option not working in proxysql cluster
General
Escalation
General
Escalation
Description
With our PXC Kubernetes operator we use --reload startup option when starting ProxySQL in case if the user sets/updates configuration through Kubernetes secret (in that case ProxySQL pods get restarted and the new config is used). This was working fine with 2.0.18, but with 2.3.2 there seems to be an issue (or we are doing something wrong).
Here's the example config:
Now if we start with this and then just update one variable "max_connections=1024" to "max_connections=4096" in config file and re-run with something like "proxysql -f -c /etc/proxysql/proxysql.cnf --reload", but the new variable is not used when we query it from admin interface.
Here's new info, we have 3 proxysql pods running in a cluster and what operator does when new config is applied it restarts one by one, but I have noticed that when I kill all proxysql pods at once then when they start they use the new config.
With our PXC Kubernetes operator we use --reload startup option when starting ProxySQL in case if the user sets/updates configuration through Kubernetes secret (in that case ProxySQL pods get restarted and the new config is used).
This was working fine with 2.0.18, but with 2.3.2 there seems to be an issue (or we are doing something wrong).
Here's the example config:
Now if we start with this and then just update one variable "max_connections=1024" to "max_connections=4096" in config file and re-run with something like "proxysql -f -c /etc/proxysql/proxysql.cnf --reload", but the new variable is not used when we query it from admin interface.
Here's new info, we have 3 proxysql pods running in a cluster and what operator does when new config is applied it restarts one by one, but I have noticed that when I kill all proxysql pods at once then when they start they use the new config.