Hi everyone!
I have a 3 Node cluster using XTraDb Cluster, 2 nodes is configurate to read and write and another node is just a galera arbiter to simulate 3 Nodes, in the node 2 have a corruption binlog problem, after that node 1 is separated to the cluster, when the service of mysql restar can’t optain the SST from node 2, any ideas about what i can do?
Files
https://github.com/KatchVargas/Problem_Percona/tree/master
It’s seems like there’s a network problem. “No route to host”
yeah! but checking with network team, we can see what the port 4444 not available in any time, the firewall and tables space is down i forgot but i use Centos 7
Test if pinging from the donor to the joiner and vice versa works. Port 4444 is not “always” on listen-mode and would only open when an SST request is received.
The ping has no problem, because they are in the same VCenter, the firewall was turned off on both servers and also the table space, it connects correctly by telnet between both servers througth port 22, as by ssh not have problems, it was done a network trace and that was when the network team realized thet port 4444 was not enabled when MySQL service restarted.