I have a NAS server running tgtd on Centos 6. In my /var/log/messages, I an getting a constant entry every 3 seconds:
Sep 26 22:21:24 san2 tgtd: conn_close(100) connection closed, 0x1143728 1 Sep 26 22:21:27 san2 tgtd: conn_close(100) connection closed, 0x10b93e8 1 Sep 26 22:21:30 san2 tgtd: conn_close(100) connection closed, 0x10b93e8 1 Sep 26 22:21:33 san2 tgtd: conn_close(100) connection closed, 0x10b93e8 1
And the tgtd service is consuming about 30% of my CPU on the SAN server, with a very light client load (only 2 clients connected, not using their disks much)
The odd thing is I don't see any problems on the client side – they connect to the iscsi target and appear to be working fine.
My targets are setup as image files without login security (just using IP address to restrict access to targets)
<target iqn.2011-0926.com.hostname.san2:ded38.target1>
backing-store /dedvol/servers/ded38.img
initiator-address 10.1.1.138
</target>
The clients are all Centos 5.6 (both x64 and i386) and the iscsi-initiator-utils is running the latest version. The NAS server is CentOS 6 x64, running the latest tgtd service.
Any ideas as to what is causing this?
Best Answer
The version of tgtd that shipped with Centos 6 (even fully patched) has this problem. I ended up having to build tgtd from source. The version I used was fujita-tgt-v1.0.27. I believe it came off the Fedora 17 site at the time, though there is probably a newer version out there now.