Unable to add iSCSI LUN

F

FORCED-INDUCTN

Member
May 5, 2010
44
0
6
California, USA
Hello!

I am having a rather annoying problem with iSCSI. I have a SAN with 2x 10GbE interfaces bonded into a 20GbE ether-channel between my SAN and my proxmox server. The SAN is running iSCSI and providing a LUN to the proxmox server for VM storage. However the SAN is in multiple subnets and it appears to advertise the LUN multiple times (one for each subnet) which in turn makes the proxmox server unable to add the iSCSI LUN (Most of the subnets are unreachable from the proxmox server, plus I want the traffic to go over the 20Gig) . If I force iSCSI to listen on a single interface it works just fine. However I now have to add another server to the SAN. And I am back to the same spot, I cant add the iSCSI LUN

On a side note, I am doing all of this so I can cluster the proxmox servers (and migrate VMs). Do I have to add the iSCSI storage to both servers or just to one, then cluster the servers? I don't remember, it has been a while since I did this last.

Both servers are running the same version of software:

FOUNDRY:~# pveversion -v
pve-manager: 1.8-18 (pve-manager/1.8/6070)
running kernel: 2.6.35-1-pve
proxmox-ve-2.6.35: 1.8-11
pve-kernel-2.6.35-1-pve: 2.6.35-11
pve-kernel-2.6.24-10-pve: 2.6.24-21
pve-kernel-2.6.18-2-pve: 2.6.18-5
qemu-server: 1.1-30
pve-firmware: 1.0-11
libpve-storage-perl: 1.0-17
vncterm: 0.9-2
vzctl: 3.0.28-1pve1
vzdump: 1.2-14
vzprocps: 2.0.11-2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.14.1-1
ksm-control-daemon: 1.0-6



Any ideas?

Thanks in advance!

--Forced
 
Last edited:
I don't think we are hitting the limit (not sure thought) because I have only been doing testing on it. But having the extra room is nice especially when I start to p2v some of my file servers. However I don't think the bonding has anything to do with the iSCSI problems, it seems to be an issue with the iSCSI target software.

I will go run a iperf on it.

*EDIT* iperf gives me:
[ 5] 0.0-16.0 sec 10.0 GBytes 5.37 Gbits/sec

--Forced

**EDIT2**
Hear is an example of what I am talking about. When I run iscsiadm on the proxmox server to manually list the iSCSI LUNs on the SAN this is what I get:
FOUNDRY:~# iscsiadm -m discovery -t st -p 192.168.66.1
192.168.69.1:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.2:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.66.5:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.8:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.66.1:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.14:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.5:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
10.1.1.41:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.11:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0

However only the 5th one from the top is relevant. So what I am wondering is, how do I tell the iSCSI target not to advertise that LUN for all those networks.
 
Last edited:
FORCED-INDUCTN said:
I don't think we are hitting the limit (not sure thought) because I have only been doing testing on it. But having the extra room is nice especially when I start to p2v some of my file servers. However I don't think the bonding has anything to do with the iSCSI problems, it seems to be an issue with the iSCSI target software.

I will go run a iperf on it.

*EDIT* iperf gives me:
[ 5] 0.0-16.0 sec 10.0 GBytes 5.37 Gbits/sec

--Forced

**EDIT2**
Hear is an example of what I am talking about. When I run iscsiadm on the proxmox server to manually list the iSCSI LUNs on the SAN this is what I get:
FOUNDRY:~# iscsiadm -m discovery -t st -p 192.168.66.1
192.168.69.1:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.2:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.66.5:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.8:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.66.1:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.14:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.5:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
10.1.1.41:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0
192.168.69.11:3260,1 vmstorage.2011-07.fortress.seq.org:iscsi.lun0

However only the 5th one from the top is relevant. So what I am wondering is, how do I tell the iSCSI target not to advertise that LUN for all those networks.
Click to expand...
Hi,
if you can't change your settings on the iscsi-target, why you dont change the routing on the pve-hosts like this:
Code: 
ip route add 192.168.69.0/24 via 192.168.66.1
But i don't understand that only the 5th is relevant - what's about 192.168.66.1? Its the same network or do you have an special network mask?

Udo
 
Yep I have a /30 mask on those interfaces, they are all point-to-point to the servers. However I also emailed the iscsi-target mailing list and I got this reply:

"You can use the /etc/iet/targets.allow file to control which interfaces a target is exposed on. By default all interfaces are utilized."

And indeed you can, I edited that file and now I have filtered out *most* of the other LUNs but of course I am stuck with the two from the two functioning interfaces. So I am going to try you route idea and just tell the servers that the other network is available over that link.

Some on else also made a good point: "it is standard practice to have a reverse FQDN in your iSCSI share,
whereas you have the regular FQDN."


I'll post back with results, thank you so much for the help!

--Forced
 
You must log in or register to reply here.
Top Bottom