-3

I'm trying to deploy a simple GlusterFS cluster on CentOS using Virtualbox. I only have two nodes (gluster-node1 and gluster-node4) with two bricks each and I'm creating a volume with replica level 2.

When I issue the volume create command on gluster-node4 I get this error: Host gluster-node1.test not connected. The glustersd daemon is running on both host and if I check the peer status it returns State: Peer in Cluster (Disconnected) on both host. I checked the /var/log/glusterfs/etc-glusterfs-glusterd.vol.log file on gluster-node4 and found this error E [socket.c:2157:socket_connect_finish] 0-management: connection to 192.168.1.61:24007 failed (No route to host), but if I try to ping or ssh to the gluster-node1 host it works without any problem.

As far as I can see there is no rule in iptables that would block this connection and I can see a route in the routing table.

# iptables --list -n
Chain INPUT (policy ACCEPT)
target     prot opt source               destination         
ACCEPT     all  --  0.0.0.0/0            0.0.0.0/0           state RELATED,ESTABLISHED 
ACCEPT     icmp --  0.0.0.0/0            0.0.0.0/0           
ACCEPT     all  --  0.0.0.0/0            0.0.0.0/0           
ACCEPT     tcp  --  0.0.0.0/0            0.0.0.0/0           state NEW tcp dpt:22 
REJECT     all  --  0.0.0.0/0            0.0.0.0/0           reject-with icmp-host-prohibited 

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination         
REJECT     all  --  0.0.0.0/0            0.0.0.0/0           reject-with icmp-host-prohibited 

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination

# route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
192.168.1.0     0.0.0.0         255.255.255.0   U     0      0        0 eth1
10.0.2.0        0.0.0.0         255.255.255.0   U     0      0        0 eth0
169.254.0.0     0.0.0.0         255.255.0.0     U     1002   0        0 eth0
169.254.0.0     0.0.0.0         255.255.0.0     U     1003   0        0 eth1
0.0.0.0         10.0.2.2        0.0.0.0         UG    0      0        0 eth0

(eth1 is an Host-only interface on the Vbox VMs)

Am I missing something? Any suggestions?

-- UPDATE --

Both peers have entries in the /etc/hosts

# more /etc/hosts (on gluster-node1)
127.0.0.1   gluster-node1.test gluster-node1 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6
192.168.1.64    gluster-node4 gluster-node4.test

# more /etc/hosts (on gluster-node4)
127.0.0.1   gluster-node4.test gluster-node4 localhost localhost.localdomain localhost4 localhost4.localdomain4
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6
192.168.1.61    gluster-node1 gluster-node1.test
dstroppa
  • 1
  • 1
  • 3

2 Answers2

0

Please check the name resultion. Does ping work with the peer names? GlusterFS tries to connect over the names, detected on peering to the other the peers.

To avoid problems due DNS, you should add some entries to /etc/hosts

Thomas Berger
  • 1,700
  • 13
  • 22
  • I thought of that too, but I've have the entries in the `/etc/hosts` for both my peers and `ping` works both using the hostname and the FQDN. `# ping gluster-node1 PING gluster-node1 (192.168.1.61) 56(84) bytes of data. 64 bytes from gluster-node1 (192.168.1.61): icmp_seq=1 ttl=64 time=0.267 ms` – dstroppa Aug 16 '13 at 11:48
  • Could you ping `gluster-node1.test` ? As it is the host in the error message. – Thomas Berger Aug 16 '13 at 11:58
  • `ping`works in all three cases: IP address, hostname and FQDN. `# ping gluster-node1 PING gluster-node1 (192.168.1.61) 56(84) bytes of data. 64 bytes from gluster-node1 (192.168.1.61): icmp_seq=1 ttl=64 time=0.118 ms` `# ping gluster-node1.test PING gluster-node1 (192.168.1.61) 56(84) bytes of data. 64 bytes from gluster-node1 (192.168.1.61): icmp_seq=1 ttl=64 time=0.052 ms` – dstroppa Aug 16 '13 at 12:07
  • Very strange. No more ideas. Maybe you should try to capture the traffic with tcpdump. – Thomas Berger Aug 16 '13 at 12:11
-1

I actually solved this issue by disabling the iptables service as described here.

dstroppa
  • 1
  • 1
  • 3