Percona XtraDB集群首次启动时间等待

Question

I am trying to start the cluster on three clean centos machines.

I tried to keep this post short, I am not attaching config files becouse I used this guide and the config files are according this:

https://www.percona.com/doc/percona-xtradb-cluster/5.7/add-node.html#add-node

Starting first node ok. Starting second node error.

Here is the log on second node

2017-09-28T15:05:09.367856Z 0 [Note] WSREP: Initiating SST/IST transfer on JOINER side (wsrep_sst_xtrabackup-v2 --role 'joiner' --address '192.168.14.104' --datadir '/var/lib/mysql/' --defaults-file '/etc/my.cnf' --defaults-group-suffix '' --parent '5490' '' ) 2017-09-28T15:05:09.368984Z 0 [ERROR] WSREP: Failed to read 'ready ' from: wsrep_sst_xtrabackup-v2 --role 'joiner' --address '192.168.14.104' --datadir '/var/lib/mysql/' --defaults-file '/etc/my.cnf' --defaults-group-suffix '' --parent '5490' '' Read: '(null)' 2017-09-28T15:05:09.369064Z 0 [ERROR] WSREP: Process completed with error: wsrep_sst_xtrabackup-v2 --role 'joiner' --address '192.168.14.104' --datadir '/var/lib/mysql/' --defaults-file '/etc/my.cnf' --defaults-group-suffix '' --parent '5490' '' : 2 (No such file or directory) 2017-09-28T15:05:09.370161Z 2 [ERROR] WSREP: Failed to prepare for 'xtrabackup-v2' SST. Unrecoverable. 2017-09-28T15:05:09.370192Z 2 [ERROR] Aborting

Answer 1

Second node startup is failing because it is unable to perform and SST (full state transfer) from the donor node.

This is failing because the xtrabackup-v2 is failing. You need to check the logs on the donor node to get more idea as to why, but possible reasons include -

• Insufficient memory on donor node
• Syntax error in my.cnf on donor node (xtrabackup is more picky about syntax than normal mysql -- check for duplicate lines, which mysql accepts but xtrabackup doesnt)
• File permissions
• xtrabackup installed incorrectly, not installed, or wrong version
• mismatch in wsrep configuration between nodes
• invalid credentials for wsrep authentication

Answer 2

There are several reasons the SST could have failed. You need to examine the logs on the first node too. Could be ports blocked, could be no SST user created, wrong SST password, missing xtrabackup software, etc, etc. Impossible to tell from only what you provided.