Starport AoE initiator with linux vblade and NTFS - corrupt

Initiator (iSCSI, FCoE, AoE, iSER and NVMe over Fabrics), iSCSI accelerator and RAM disk

Moderators: art (staff), anton (staff), Anatoly (staff), Max (staff)

Post Reply
dmd350
Posts: 3
Joined: Wed Jan 20, 2010 11:01 pm

Wed Jan 20, 2010 11:16 pm

I'm using the Starport AoE initiator on 4 servers to an Ubuntu 9.10 linux server target running vblade on an NTFS partition. I'm not sure what I'm doing wrong, but data on the target is being corrupted in strange and mysterious ways when I write to the volume from each of the different systems. For example, If I copy folder A from server 1, then copy folder B from server 2, then copy folder C from server 3. All servers should see folders A, B, and C, but instead each can only see the 1 folder that it copied. On the vblade target server, only folder C is visible. Each server can browse the contents of the folder it copied until it reboots, then it says "File system is corrupt and unusable."

I thought NTFS was supposed to be able to deal with write access from multiple systems, but it clearly isn't. So what file system should be used on the target?

Thanks,
DD
dmd350
Posts: 3
Joined: Wed Jan 20, 2010 11:01 pm

Thu Jan 21, 2010 12:53 am

I figured it out. Thought I was being smart by utilizing resources the way I did, but it didn't work out like I thought it would. I'm posting this solution for the benefit of anyone else who may also have this problem out there.

Each of my servers has 8 gigE ports. I thought by eliminating the switch and connecting these servers to each other directly on separate ports at the target I would reduce latency, eliminate collisions and increase throughput. I did this by starting a separate instance of vblade for each interface on the target, meaning eth1-eth4 had a vblade daemon running on each. That was the source of my problem. I reconfigured it to use 1 vblade daemon on 1 interface using a gigE switch, and disabled all write caching on the initiators. That was the end of it.
Constantin (staff)

Thu Jan 21, 2010 8:59 am

Good to hear that everything is OK now.
Post Reply