[Gluster-users] distributed geo-rep using hostname instead of FQDN of slave node

[ML mail]

Hi,

I just setup distributed geo replication on my two nodes replica master glusterfs 3.7.11 towards my single slave node replica and noticed that for some reasons it takes the hostname of my slave node instead of the fully qualified domain name (FQDN) and this although I have specified the FQDN in the gluster volume geo-replication create command. As my slave node is on another network with another domain name geo replication does not work unless I manually add the hostname to all my master nodes /etc/hosts file. 


As an example this is the type of error you get due to this issue (taken from the 
/var/log/glusterfs/geo-replication/myvolume/ssh%3A%2F%2Froot%40192.168.100.101%3Agluster%3A%2F%2F127.0.0.1%3Amyvolume-geo.log log file):


[2016-06-25 21:57:38.964888] E [resource(/data/myvolume/brick):230:logerr] Popen: ssh> ssh: Could not resolve hostname mygeohost: Name or service not known

I guess this is a bug in the geo replication create scripts or code and should be addressed.

Best regards,
ML