I think an update from 10.1.14 to 10.1.16 might be the culprit?
I try to restart the cluster and it fails:
in logs:

2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Compressed tables use zlib 1.2.7
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Using Linux native AIO
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Using SSE crc32 instructions
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Initializing buffer pool, size = 128.0M
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Completed initialization of buffer pool
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Highest supported file format is Barracuda.
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: 128 rollback segment(s) are active.
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Waiting for purge to start
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Percona XtraDB (http://www.percona.com) 5.6.30-76.3 started; log sequence number 110576354
2016-07-26 14:48:15 139717081184384 [Warning] InnoDB: Skipping buffer pool dump/restore during wsrep recovery.
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Creating #1 thread id 139716292761344 total threads 4.
2016-07-26 14:48:15 139717081184384 [Note] InnoDB: Creating #2 thread id 139716275984128 total threads 4.
2016-07-26 14:48:16 139717081184384 [Note] InnoDB: Creating #3 thread id 139716267591424 total threads 4.
2016-07-26 14:48:16 139717081184384 [Note] InnoDB: Creating #4 thread id 139716259198720 total threads 4.
2016-07-26 14:48:16 139717081184384 [Note] Plugin 'FEEDBACK' is disabled.
2016-07-26 14:48:16 139717081184384 [Note] Server socket created on IP: '::'.
2016-07-26 14:48:16 139717081184384 [Note] WSREP: Recovered position: cc2ae15e-fa6b-11e5-aaba-da42ff198f35:2225

at the same time in systemctl:
16455]: WSREP: Failed to recover position: '2016-07-26 14:48:15 139717081184384 [Note] /usr/sbin/mysqld (mysqld 10.1.16-MariaDB) starting as process 16533 ...'

regards.