garbd does not work when specifying cluster address without the port

[Jaime Sicam]

Tested on: Percona-XtraDB-Cluster-garbd-3.x86_64 0:3.7-1.3254.rhel6

If I try running this to join cluster without specifying port 4567:
garbd -a gcomm://192.168.1.90 -g my_centos_cluster
2014-09-03 23:45:04.669 INFO: CRC-32C: using "slicing-by-8" algorithm.
2014-09-03 23:45:04.669 INFO: Read config:
 daemon: 0
 name: garb
 address: gcomm://192.168.1.90
 group: my_centos_cluster
 sst: trivial
 donor:
 options: gcs.fc_limit=9999999; gcs.fc_factor=1.0; gcs.fc_master_slave=yes
 cfg:
 log:

2014-09-03 23:45:04.671 INFO: protonet asio version 0
2014-09-03 23:45:04.671 INFO: Using CRC-32C for message checksums.
2014-09-03 23:45:04.671 INFO: backend: asio
2014-09-03 23:45:04.672 WARN: access file(gvwstate.dat) failed(No such file or directory)
2014-09-03 23:45:04.672 INFO: restore pc from disk failed
2014-09-03 23:45:04.673 INFO: GMCast version 0
terminate called after throwing an instance of 'gu::NotSet'
Aborted

If I specify the port, it works:
 garbd -a gcomm://192.168.1.90:4567 -g my_centos_cluster
2014-09-03 23:46:21.084 INFO: CRC-32C: using "slicing-by-8" algorithm.
2014-09-03 23:46:21.084 INFO: Read config:
 daemon: 0
 name: garb
 address: gcomm://192.168.1.90:4567
 group: my_centos_cluster
 sst: trivial
 donor:
 options: gcs.fc_limit=9999999; gcs.fc_factor=1.0; gcs.fc_master_slave=yes
 cfg:
 log:

2014-09-03 23:46:21.086 INFO: protonet asio version 0
2014-09-03 23:46:21.086 INFO: Using CRC-32C for message checksums.
2014-09-03 23:46:21.086 INFO: backend: asio
2014-09-03 23:46:21.087 WARN: access file(gvwstate.dat) failed(No such file or directory)
2014-09-03 23:46:21.087 INFO: restore pc from disk failed
2014-09-03 23:46:21.088 INFO: GMCast version 0
2014-09-03 23:46:21.088 INFO: (73949133, 'tcp://0.0.0.0:4567') listening at tcp://0.0.0.0:4567
2014-09-03 23:46:21.088 INFO: (73949133, 'tcp://0.0.0.0:4567') multicast: , ttl: 1
2014-09-03 23:46:21.089 INFO: EVS version 0

However, if you check default config on /etc/sysconfig/garb, it shows that specifying the port is optional:
cat /etc/sysconfig/garb
# Copyright (C) 2012 Coedership Oy
# This config file is to be sourced by garb service script.

# REMOVE THIS AFTER CONFIGURATION

# A space-separated list of node addresses (address[:port]) in the cluster
# GALERA_NODES=""

# Galera cluster name, should be the same as on the rest of the nodes.
# GALERA_GROUP=""

# Optional Galera internal options string (e.g. SSL settings)
# see http://www.codership.com/wiki/doku.php?id=galera_parameters
# GALERA_OPTIONS=""

# Log file for garbd. Optional, by default logs to syslog
# LOG_FILE=""

But say if you specify the cluster address without the ports and start garbd:
Eg. GALERA_NODES="192.168.1.90 192.168.1.91"

You get the error:
[root@localhost ~]# service garb start
nc: port range not valid
nc: port range not valid
nc: port range not valid
None of the nodes in 192.168.1.90 192.168.1.91 is accessibl[FAILED]

My suggestion is to either make port 4567 default for garbd or change "# A space-separated list of node addresses (address[:port]) in the cluster" to "# A space-separated list of node addresses (address:port) in the cluster"

[Przemek (pmalkowski)]

I confirm that for both Percona-XtraDB-Cluster-garbd-2-2.11-1.2675.rhel6 and Percona-XtraDB-Cluster-garbd-3-3.7-1.3254.rhel6.
I agree the port should be either mandatory or 4567 set as default.

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PXC-1729