Redis-trib by using hostnames fails to join cluster

[andrepintorj]

I was trying to setup a redis cluster by using hostnames instead of ip addresses as follows:

[root@poll-a bin]# ./redis-trib.rb create poll-a.mycompany.com:7000 poll-b.mycompany.com:7000 poll-c.mycompany.com:7000

Creating cluster
Connecting to node poll-a.mycompany.com:7000: OK
Connecting to node poll-b.mycompany.com:7000: OK
Connecting to node poll-c.mycompany.com:7000: OK
Performing hash slots allocation on 3 nodes...
Using 3 masters:
poll-a.mycompany.com:7000
poll-b.mycompany.com:7000
poll-c.mycompany.com:7000
M: 095321f50fd225c1834bc9b319f6dc8ea92583fd poll-a.mycompany.com:7000
slots:0-5460 (5461 slots) master
M: 0e694d3eff8c70d2076f00f11d8b6bef4aeb94f0 poll-b.mycompany.com:7000
slots:5461-10922 (5462 slots) master
M: 2758cf90c9141cb647fdff20aabf829e63504700 poll-c.mycompany.com:7000
slots:10923-16383 (5461 slots) master
Can I set the above configuration? (type 'yes' to accept): yes
Nodes configuration updated
Assign a different config epoch to each node
Sending CLUSTER MEET messages to join the cluster
/usr/local/rvm/gems/ruby-2.1.0/gems/redis-3.1.0/lib/redis/client.rb:103:in call': ERR Invalid node address specified: poll-a.mycompany.com:7000 (Redis::CommandError) from /usr/local/rvm/gems/ruby-2.1.0/gems/redis-3.1.0/lib/redis.rb:2486:inblock in method_missing'
from /usr/local/rvm/gems/ruby-2.1.0/gems/redis-3.1.0/lib/redis.rb:37:in block in synchronize' from /usr/local/rvm/rubies/ruby-2.1.0/lib/ruby/2.1.0/monitor.rb:211:inmon_synchronize'
from /usr/local/rvm/gems/ruby-2.1.0/gems/redis-3.1.0/lib/redis.rb:37:in synchronize' from /usr/local/rvm/gems/ruby-2.1.0/gems/redis-3.1.0/lib/redis.rb:2485:inmethod_missing'
from ./redis-trib.rb:664:in block in join_cluster' from ./redis-trib.rb:662:ineach'
from ./redis-trib.rb:662:in join_cluster' from ./redis-trib.rb:974:increate_cluster_cmd'
from ./redis-trib.rb:1345:in `

'

Interesting to see is that it performs the node configuration and finds/connects to the hosts, but then while sending cluster MEET messages, it throws an error as shown above.

If I do exactly the same commands by using IP addresses it works flawlessly:

[root@poll-a bin]# ./redis-trib.rb create 10.0.2.101:7000 10.0.2.104:7000 10.0.2.224:7000

Creating cluster
Connecting to node 10.0.2.101:7000: OK
Connecting to node 10.0.2.104:7000: OK
Connecting to node 10.0.2.224:7000: OK
Performing hash slots allocation on 3 nodes...
Using 3 masters:
10.0.2.101:7000
10.0.2.104:7000
10.0.2.224:7000
M: 5bd2e31fef497f019ba440509ba34b264b233b7b 10.0.2.101:7000
slots:0-5460 (5461 slots) master
M: db67c6ed7b58a251b7ca6a6699d3ddfde9178182 10.0.2.104:7000
slots:5461-10922 (5462 slots) master
M: 50e3ce57a9b211c44df9aba561810b60e6e2e833 10.0.2.224:7000
slots:10923-16383 (5461 slots) master
Can I set the above configuration? (type 'yes' to accept): yes
Nodes configuration updated
Assign a different config epoch to each node
Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join....
Performing Cluster Check (using node 10.0.2.101:7000)
M: 5bd2e31fef497f019ba440509ba34b264b233b7b 10.0.2.101:7000
slots:0-5460 (5461 slots) master
M: db67c6ed7b58a251b7ca6a6699d3ddfde9178182 10.0.2.104:7000
slots:5461-10922 (5462 slots) master
M: 50e3ce57a9b211c44df9aba561810b60e6e2e833 10.0.2.224:7000
slots:10923-16383 (5461 slots) master
[OK] All nodes agree about slots configuration.
Check for open slots...
Check slots coverage...
[OK] All 16384 slots covered.

[mattsta]

Correct! Hostnames are not supported. IPs only.

[mcpherson]

I stumbled across this too. Easy enough to fix, but why no host names?

[badboy]

Well, #2323 is there...