On Oct 11, 2009, at 5:14 PM, C Linus Hicks linush@verizon.net wrote:
On Sun, 2009-10-11 at 15:13 -0400, C Linus Hicks wrote:
There's no VIF on loopback. The machine running iet has two NICs, the private interconnect should not be playing a role. It's IP address is 10.200.2.2/24 while the local subnet is 10.200.1.0/24 and the iet machine is IP address 10.200.1.6.
From running strace on the iscsid daemon, it appears to be taking what
looks to me like an unexpected error. In the text below, look for the first series of "poll" system calls. Just after that section is where I initiate the iscsiadm login command. Then about 100 lines later on, it issues a connect call that returns "-1 EINPROGRESS (Operation now in progress)". Then it goes back to polling, this time on two sockets including the one it just tried to connect on. In a little while it probably reaches the retry limit and closes that socket. It makes another connect attempt that fails with the same error. There are a total of five connect attempts, but I have cut out some of the repeated text.
I don't know much about open-iscsi, just IET and the best way to diagnose iSCSI problems is via tcpdump and wireshark not strace.
Is the iscsi-target service running?
-Ross