iSCSI mount zvol as boot device using Chelsio T520-SO-CR

[Scampicfx]

Dear folks,

setup of two servers:
- FreeNAS Server, Xeon E5, Chelsio T520-CR
- "Disk-less" VM Server, Xeon D, Chelsio T520-SO-CR

I configured the iSCSI service as mentioned in the documentation and as shown in this helpful youtube video: https://www.youtube.com/watch?v=HvyOWlFISdo

It's running fine so far. I tested to connect to the volume using win 7 client -> cool!

Now, I want the additional disk-less VM server to use a zvol of FreeNAS as boot device! This disk-less server is equipped with Chelsio T520-SO-CR. So, after configuring the chelsio network adapter of the disk-less VM server, the FreeNAS target is found. The disk-less server connects to this target. So, the next step is searching for LUNs!
However... When starting to look for LUNs, the disk-less server reports that "No LUNs [are] present on the chosen target" (target = FreeNAS).

So, after reading, reconfiguring, and so on all day long, I have no idea why the disk-less server can't find any LUNs? To be honest, I don't know if the problem is located at FreeNAS or at the Chelsio Network Adapter installed in the disk-less-server?

When looking in the logs of FreeNAS, i notice following:

Jun 23 22:25:35 storageunit notifier: Starting ctld.
Jun 23 20:31:03 storageunit ctld[80089]: 10.2.100.39 (iqn.2005-10.org.freenas.ctl): exiting due to timeout
Jun 23 22:31:03 storageunit ctld[79626]: child process 80089 terminated with exit status 1
Jun 23 22:33:58 storageunit ctld[80328]: 10.2.100.39 (iqn.2005-10.org.freenas.ctl): exiting due to timeout
Jun 23 22:33:58 storageunit ctld[79626]: child process 80328 terminated with exit status 1

Does anybody know what "exiting due to timeout" means? Is an error more likely to find at disk-less server or FreeNAS?

My goal is, that all the VM server don't need any hard drives any more and instead use FreeNAS as central data storage ;)

 

[Scampicfx]

I noticed following information in one of the network adapter manuals:

###########
To support iSCSI Boot, the target needs to support multiple sessions from the same initiator. Both the iSCSI Boot firmware initiator and the OS High Initiator need to establish an iSCSI session at the same time. Both these initiators use the same Initiator Name and IP Address to connect and access the OS disk but these two initiators will establish different iSCSI sessions. In order for the target to support iSCSI Boot, the target must be capable of supporting multiple sessions and client logins.
##########

Does FreeNAS support multiple sessions for the same target?

 

[dlavigne]

Did you find the answer to this?

 

[Scampicfx]

Not yet... I'm still in contact with Chelsio due to this issue. As far as I know Chelsio set up a similar FreeNAS environment in their labs in order to test this issue.
Meanwhile, I ordered a different NIC, since we have to get the additional disk-less server online asap.
I hope this issue can be solved!

 

[Scampicfx]

Dear guys,
I replaced Chelsio T520-SO-CR by new Intel X710.
Now, target and LUNs are found! :)

Anyway... There are still a few problems... Most probably minor problems... I hope anyone can help. Any idea or information is highly appreciated!

In my log, following messages appear when mounting the LUN:

Jul 4 23:25:50 storageunit ctld[99775]: 10.2.100.9: invalid target name "user-tractorunitboot"; should start with either "iqn.", "eui.", or "naa."
Jul 4 23:26:01 storageunit ctld[99787]: 10.2.100.9: invalid target name "user-tractorunitboot"; should start with either "iqn.", "eui.", or "naa."

My question is: What exactly is invalid?

My target is named: targettractorunitboot
My user is named: user-tractorunitboot

Intel NIC is connecting via iqn.2017-04.org.storageunit.ctl:targettractorunitboot
LUN ID is 0.

What exactly is invalid?

Second question:

As soon as there is some sort of inactivity, I receive the following messages:

Jul 4 23:26:06 storageunit WARNING: 10.2.100.9 (user-tractorunitboot): no ping reply (NOP-Out) after 5 seconds; dropping connection

Since it is classified as "WARNING", I'm a bit concerned about this issue. Why does this happen? What's the reason and how can I solve it?

As mentioned, any help (and any tiny help) are highly appreciated :)

Thanks!