I have a trio of CentOS 6.4 machines that I'm trying to configure to use iSCSI. Basically two different initiators talking to two different targets on a single host server.
On one machine, "the server", I have two targets defined in
/etc/tgt/targets.conf each target points to a different logical disk. The configuration stanzas are nearly identical, varying only in their names and
These are the lines from
/etc/tgt/targets.conf that aren't commented out.
default-driver iscsi <target iqn.2013-08.com.example.group:backup> backing-store /dev/primary/backup initiator-address <ip address> incominguser username password </target> <target iqn.2013-08.com.example.group:migrations> backing-store /dev/primary/migrations initiator-address <ip address> incominguser username password </target>
When I run
tgt-admin -s I show two targets as expected. However, one of the targets shows only 1 LUN (0 Controller) while the other target shows 2 LUNs (a Controller and a Disk). I have no idea if this is significant.
Sadly, the first target (w/o a disk LUN) is able to be reached by it's initiator and the machine can mount the space and it's all peaches and cream. The second target (w/ the disk LUN) can't be reached by the initiator. When I run
iscsiadm -m discovery -t sendtargets -p <ipaddr> I get back
iscsiadm: No portals found.
What's worse, even if I when the second target to expect connection from the other address, or remove the
initiator-address portion of the configuration, no other systems seem to see the second target.