Comment 15 for bug 1111706

I can see the same problem when, for example SST configuration is incompatible between the donor and joiner, but even in a situation where IST would be completely enough to join the cluster (tested on PXC 5.6.26 and MariaDB Cluster 10.0.21):

[root@percona2 ~]# /etc/init.d/mysql stop
Shutting down MySQL (Percona XtraDB Cluster)..... SUCCESS!

[root@percona2 ~]# cat /data/myisam/grastate.dat
# GALERA saved state
version: 2.1
uuid: fadf885a-490b-11e4-9e13-27f574d828b6
seqno: 449486
cert_index:

-- edit my.cnf on joiner only:
[sst]
#streamfmt=xbstream
streamfmt=tar

error log fragment on restart:

2015-10-24 14:43:38 19271 [Note] WSREP: Requesting state transfer: success, donor: 0
tar: This does not look like a tar archive
tar: Exiting with failure status due to previous errors
2015-10-24 14:43:40 19271 [Note] WSREP: 0.0 (percona1): State transfer to 1.0 (percona2) complete.
2015-10-24 14:43:40 19271 [Note] WSREP: Member 0.0 (percona1) synced with group.
WSREP_SST: [ERROR] Error while getting data from donor node: exit codes: 0 2 (20151024 14:43:40.673)
WSREP_SST: [ERROR] Cleanup after exit with status:32 (20151024 14:43:40.679)
2015-10-24 14:43:40 19271 [ERROR] WSREP: Process completed with error: wsrep_sst_xtrabackup-v2 --role 'joiner' --address '192.168.99.3:4444' --datadir '/data/myisam/' --defaults-file '/etc/my.cnf' --defaults-group-suffix '' --parent '19271' '' : 32 (Broken pipe)
2015-10-24 14:43:40 19271 [ERROR] WSREP: Failed to read uuid:seqno from joiner script.
2015-10-24 14:43:40 19271 [ERROR] WSREP: SST script aborted with error 32 (Broken pipe)
2015-10-24 14:43:40 19271 [ERROR] WSREP: SST failed: 32 (Broken pipe)
2015-10-24 14:43:40 19271 [ERROR] Aborting
...

[root@percona2 ~]# cat /data/myisam/grastate.dat
# GALERA saved state
version: 2.1
uuid: 00000000-0000-0000-0000-000000000000
seqno: -1
cert_index:

Should the SST misconfiguration lead to IST failure and grastate.dat reset? IMHO not.