Any node is not coming up after all nodes goes down in PXC

I am facing issue in PXC cluster Setup .

Issue Description : I have setup three nodes PXC setup and it’s working fine but if all nodes goes down or stop mysql service on all three nodes at a time . No any node is coming up and getting error in error logs like count not find primary view . So added pc.recovery options as true and it create a gvwstate file after start and this file contain primary component view but once get server reboot and service stop then this file automatically remove from server then getting error as 2018-07-27 07:12:20 1720 [Warning] WSREP: access file(/var/lib/mysql//gvwstate.dat) failed(No such file or directory)

2018-07-27 07:12:20 1720 [Note] WSREP: restore pc from disk failed

Can anyone help for this .

Please note that starting and stopping PXC nodes is very different from traditional replication nodes. Ideally, you should never stop all nodes. If you do, you have to pick up the right candidate later, which you will use for bootstrapping the cluster again. Normal “start” won’t work as it needs working Primary Component it will connect to.

Now, pc.recovery option is only for sudden whole cluster down scenarios like datacenter power outage or similar. It does not apply when you stop or kill nodes one by one.

More examples: [url]https://www.percona.com/blog/2014/09/01/galera-replication-how-to-recover-a-pxc-cluster/[/url]