After Launching Cluster Using v1.4.0, Proxysql Never Becomes Available

MySQL & MariaDB Percona Operator for MySQL
1

I created an issue on your tracker here but have yet to get any response. I’m really stuck on this and would appreciate any guidance that might help me resolve the problem. The pxc-monit container in the proxysql pod on each node never becomes available.  The logs for the container on each node look like this:

2020/05/07 20:16:15 Failed to execute /usr/bin/add_pxc_nodes.sh: + main 
+ echo 'Running /usr/bin/add_pxc_nodes.sh' 
Running /usr/bin/add_pxc_nodes.sh 
+ read -ra first_host 
+ '[' -z db-cluster-1-pxc-0.db-cluster-1-pxc.pxc.svc.cluster.local ']' 
++ echo db-cluster-1-pxc-0.db-cluster-1-pxc.pxc.svc.cluster.local 
++ cut -d . -f 2- 
+ service=db-cluster-1-pxc.pxc.svc.cluster.local 
+ sleep 15s 
+ wait_for_mysql db-cluster-1-pxc.pxc.svc.cluster.local 
+ local h=db-cluster-1-pxc.pxc.svc.cluster.local 
+ echo 'Waiting for host db-cluster-1-pxc.pxc.svc.cluster.local to be online...' 
Waiting for host db-cluster-1-pxc.pxc.svc.cluster.local to be online... 
++ mysql_root_exec db-cluster-1-pxc.pxc.svc.cluster.local 'select 1' 
++ local server=db-cluster-1-pxc.pxc.svc.cluster.local 
++ local 'query=select 1' 
++ MYSQL_PWD=1KO5UIduAdcQscyt 
++ timeout 600 mysql -h db-cluster-1-pxc.pxc.svc.cluster.local -uroot -s -NB -e 'select 1' 
+ '[' 1 '!=' 1 ']' 
+ wait_for_proxy 
+ local h=127.0.0.1 
+ echo 'Waiting for host 127.0.0.1 to be online...' 
Waiting for host 127.0.0.1 to be online... 
++ proxysql_admin_exec 127.0.0.1 'select 1' 
++ local server=127.0.0.1 
++ local 'query=select 1' 
++ MYSQL_PWD=3BhMOqF5A521fy2Y 
++ timeout 600 mysql -h 127.0.0.1 -P6032 -u proxyadmin -s -NB -e 'select 1' 
+ '[' 1 '!=' 1 ']' 
+ SSL_ARG= 
++ proxysql_admin_exec 127.0.0.1 'SELECT variable_value FROM global_variables WHERE variable_name="mysql-have_ssl"' 
++ local server=127.0.0.1 
++ local 'query=SELECT variable_value FROM global_variables WHERE variable_name="mysql-have_ssl"' 
++ MYSQL_PWD=3BhMOqF5A521fy2Y 
++ timeout 600 mysql -h 127.0.0.1 -P6032 -u proxyadmin -s -NB -e 'SELECT variable_value FROM global_variables WHERE variable_name="mysql-have_ssl"' 
+ '[' true = true ']' 
+ SSL_ARG=--use-ssl=yes 
+ proxysql-admin --config-file=/etc/proxysql-admin.cnf --cluster-hostname=db-cluster-1-pxc-0.db-cluster-1-pxc.pxc.svc.cluster.local --enable --update-cluster --remove-all-servers --disable-updates --force --debug --use-ssl=yes 
debug (line:2771) : Loading /etc/proxysql-admin.cnf 
debug: ProxySQL datadir: /var/lib/proxysql 
debug (line:712) : exec_sql : proxyadmin@localhost:6032 ==> SELECT 1 
mysql: [ERROR] Could not open required defaults file: /dev/stdin 
mysql: [ERROR] Fatal error in defaults handling. Program aborted! 
debug: --> query failed 1 
ERROR (line:713) : ProxySQL connection check failed. 
-- Could not connect to ProxySQL at localhost:6032 
Smart Checklist
2

Your Jira issue is in our queue.  Sorry we haven’t gotten to it yet.  It should be soon.

3


4

At this point I have to ask, has anyone successfully installed a working cluster using v.1.4.0 of the K8s Operator? 

5

Yes, me)
Just followed the standard procedure listed in docs with GKE platform.

6

I was able to identify a fix. See the jira ticket linked in my first post for details.

7

I was able to identify a fix. See the jira ticket linked in my first post for details.