Percona 5.7.22-22, Ubuntu 18.04 LTS - Cluster restart on Reboot without bootstrapping

mysql> SHOW VARIABLES LIKE ‘%version%’;
±------------------------±------------------------------------------------------------------------------------------------+
| Variable_name | Value |
±------------------------±------------------------------------------------------------------------------------------------+
| innodb_version | 5.7.22-22 |
| protocol_version | 10 |
| slave_type_conversions | |
| tls_version | TLSv1,TLSv1.1,TLSv1.2 |
| version | 5.7.22-22-57-log |
| version_comment | Percona XtraDB Cluster (GPL), Release rel22, Revision da86071, WSREP version 29.26, wsrep_29.26 |
| version_compile_machine | x86_64 |
| version_compile_os | debian-linux-gnu |
| version_suffix | -57-log |
±------------------------±------------------------------------------------------------------------------------------------+
9 rows in set (0.01 sec)

I am running a three node cluster. Everything is working great, but for some reason, I cannot get my cluster to come back online after rebooting the cluster stack of three servers.

I followed this article:

[url]https://www.percona.com/blog/2014/12/03/auto-bootstrapping-an-all-down-cluster-with-percona-xtradb-cluster/[/url]

and was able to bring up the cluster after doining a killall -9 mysqld on all of my nodes, came right back up just as it should.

When I reboot my cluster, I can go to my primary node (what I call primary), bootstrap it (/etc/init.d/mysql restart-bootstrap) then start all other nodes and everything comes up fine.

I guess my question is simple - how to I make sure the cluster comes back online unattended after a cluster reboot?

Many Thanks

If you only reboot 1 at a time you should not have this problem?

So the fix is not to reboot all of the cluster at the same time? I am more concerned about something like a power outage that takes my cluster down.

I am trying to determine why the cluster does not come back online without having to bootstrap it at all.

Is this not possible?

If all 3 go down at once, then you have to pick which one you like as the bootstrap node . Start up that node regularly and then run:

SET GLOBAL wsrep_provider_options=‘pc.bootstrap=true’;

Then you can turn on the other 2.

After they all sync up, I don’t know for certain if you need to then set this back to false - maybe someone else can chime in.
SET GLOBAL wsrep_provider_options=‘pc.bootstrap=false’;

Got it, so if they all fail for some reason, or if you shut down all three nodes - manual intervention is required to bring the cluster back online. That is what I needed to know. Thank You for the help!