Yet another timed out connection

[Mauricio]

I know this has been asked many time before but I still do not know what I am doing wrong. I am handing out iSCSI LUNs from a host at 192.168.10.18:3260 to a host called testbox (initiator). 

[root@testbox ~]# iscsiadm -m discovery -t sendtargets -p 192.168.10.18

192.168.10.18:3260,1 iqn.2000-01.com.synology-iSCSI:storage.01

[fe80::211:32ff:fe15:74eb]:3260,1 iqn.2000-01.com.synology-iSCSI:storage.01

[root@testbox ~]#

[root@testbox ~]# fgrep address /var/lib/iscsi/nodes/iqn.2000-01.com.synology-iSCSI\:storage.01/192.168.10.18\,3260\,1/default

node.discovery_address = 192.168.10.18

node.conn[0].address = 192.168.10.18

[root@testbox ~]#

When I try to connect I am getting the connection timed out issue. Correct me if I am wrong but it is barking at when It tries to connect using IPv6:

[root@testbox ~]# iscsiadm -m node --loginall all

Logging in to [iface: default, target: iqn.2000-01.com.synology-iSCSI:storage.01, portal: 192.168.10.18,3260]

Logging in to [iface: default, target: iqn.2000-01.com.synology-iSCSI:storage.01, portal: fe80::211:32ff:fe15:74eb,3260]

Login to [iface: default, target: iqn.2000-01.com.synology-iSCSI:storage.01, portal: 192.168.10.18,3260] successful.

iscsiadm: Could not login to [iface: default, target: iqn.2000-01.com.synology-iSCSI:storage.01, portal: fe80::211:32ff:fe15:74eb,3260].

iscsiadm: initiator reported error (8 - connection timed out)

iscsiadm: Could not log into all portals

[root@testbox ~]#

which sometimes seems to be what it wants to do by default:

[root@testbox ~]# iscsiadm -m node -T iqn.2000-01.com.synology-iSCSI:storage.01 -l

Logging in to [iface: default, target: iqn.2000-01.com.synology-iSCSI:storage.01, portal: fe80::211:32ff:fe15:74eb,3260]

iscsiadm: Could not login to [iface: default, target: iqn.2000-01.com.synology-iSCSI:storage.01, portal: fe80::211:32ff:fe15:74eb,3260].

iscsiadm: initiator reported error (8 - connection timed out)

iscsiadm: Could not log into all portals

[root@testbox ~]#

I did not really setup IPv6 in this network; is I guesstimation for the source of the problem correct? 

[The Lee-Man]

The target thinks there are two ways to reach it, so during discovery it is telling you both of those paths. Discovery, by design, reports all LUNs on the target, not just the one you want.

So now, after discovery, your iscsi database thinks there are two IQNs to connect to, both reaching the same exact target. so when you then tell iscsi to connect to that target, it tries to connect to that target through all the paths it has -- and there are two of them. But it looks like your target is actually not reachable on the IPv6 Path, for some reason. Perhaps the ACL you've set up on your target? The actual reason doesn't matter to the initiator -- it just tries and fails to talk to the target through the IPv6 path.

There are a couple of ways you can fix this. After discovery, you can delete the database node you do not want. Simply run "iscsi -m node" to see the list of database nodes discovered, then run 'iscsiadm -m node --op delete -p "fe80::211:32ff:fe15:74eb,3260]"' to delete the IPv6 node. Another option would be to leave the nodes in the database, but specify the node you wish to use when logging on. (Right now, you're telling it to login to all nodes in the database). If you still want IPv6 running on your box, you could fix the issue with your target so that open-iscsi could log into both the IPv4 and IPv6 nodes, as it is trying to do.