Strange iSCSI errors

[derwho]

Hi!

I've got an environment with 4 Dell 1955 blades connected thru VMware iSCSI software initiator to a EMC CX3-10c SAN.

All works fine except I get a shitload of the following in my /var/log/messages:

\--\[CLiP]--

Jun 6 14:20:07 esx2 iscsid\[1899]: failed to recv a PDU response, reconnecting

Jun 6 14:20:18 esx2 iscsid\[1899]: short PDU header read from socket 0: Connecti

on reset by peer

\--\[CLiP]--

From a forum post I found out that setting "Continuous=no" in my /etc/iscsi.conf might help, but unfortunately it didn't.

Any help is appriciated.

Best regards,

Anssi from Helsinki

[Paul_Lalonde]

I believe this is a known issue with the older iSCSI initiator codebase being used. I have also seen similar error messages when using different iSCSI targets that the PDU size exceeded a certain buffer limit so the connection was being disconnected. It would get rebuilt again immediately.

Hopefully these minor issues will be resolved in the next update to ESX.

Paul

[gjlowe]

I am seeing this exact same issue. However, in my case, this had been working just fine. Now, when we try to rescan the iscsi software initiator, it finds no LUNs and we see this error in the logs.

[David_N]

I am also getting this same message on a CX3-10c. I'm at the latest patch level and am seeing these messages approximately every 10 seconds.

[femialpha]

I believe this has to do with the target discovery option for the software initiator. Open the file /etc/vmkiscsi.conf and search for the line 'continuous=yes' and change it to no. There are quite a number of posts in the forum about the issue.

[David_N]

Thanks - Yeah, that was definitely it.

http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&docType=kc&externalId=3344257