I installed ESXI last night and plan to use it for creating a test environment. I have spent about 6 hours now trying to get the iSCSI connection setup. The target is on an HP AIO 600. The array is inside of an HP MSA 70 connected to the HP AIO 600 via a P800 controller card.
I have set the target up with no authentication. I am able to connect to the target from another Windows server. The Server running ESXI only has 1 network card. From within Infrastucture Client, it never sees any targets.
I checked the event logs on the Target machine and I see where the initiator logged in at one point, but I never see any targets.
Any advice?
I decided to download and install Starwind iSCSI target on my workstation and create a target to see if my ESXI would be able to connect to it, and IT DID! Not sure where to go from here. I'm glad to see that iSCSI is working on the ESXI box, but I need it to work with my HP AIO setup.
As I recall there were a few issues when I first set it up. I remember a reference to ESXi not liking LUN 0 and the second was that I had to mount the drive with some other tool first. In my case I added it to an XP workstation. I don't beleive I did much other than use disk manager to initialize it. I then disconnected it from XP and mounted it to ESXi.
Changes to iSCSI on the ESXi side require a reboot of the server so get it right before you get to production.
You can use the MS iscsi initiator in Windows.
I created two LUNs on my SAN and connected them to another server using the Microsoft ISCSI initiator. At this point I rescanned from within ESXi and it did not find the target. At that point I logged off the iSCSI connection from the other server and rescanned from ESXi and still no target. One thing I did notice was that the ISCSI connection I created between ESXI and my workstation (Using Starwind) is called "SCSI Target 2". I found this interesting because it is the only target showing in ESXi.
Any other suggestions. I'm going to reboot my server again (I don't think I've made any changes since the last reboot, but maybe I'm mistaken.
Upon a server reboot and rescan of the targets. The above mentioned "SCSI Target 2" is now SCSI Target 4. My guess is the ESXI is definitely seeing the luns I created on the AIO/MSA, but not connecting to them completely or something??????
According to the iSCSI target, the initiator is "logged in".
Look at the ESX logs and see what messages are shown after you issue a rescan task.
-KjB
VMware vExpert
iSCSI: session 0x352180a0 connect timed out at 4939130
iSCSI: 0x352180a0 retrying all the portals again, since the portal list got exhausted
iSCSI: session 0x352180a0 to "name of my target" waiting 60 seconds before next login attempt
One line a little further down that I thought was interesting was this:
iSCSI: bus 0 target 3 trying to establish session 0x352402x0 to portal 0, address 10.106.1.11 port 3260 group 1
10.106.1.11/16 is one of the IP Adresses for our SAN, but it is on a different Subnet. On the Dynamic Discovery tab I have entered 10.104.7.15 as the address to our SAN. Why is it no trying to connect through the 10.104.7.15 address instead of the 10.106.1.11? I'm using Microsoft iSCSI target software.
Thanks!
When you had the MS initiator connected did you initialize the disk? You don't need to partition just initialize.
I initialized and partitioned one of the disks and just initialized the other.
Are the subnets isolated? Subnet mask? If the SAN is connected to both subnets if I remember is can announce both addresses.
Yes, the subnets are isolated and I can't allow routing between then. Is there any way to force it to use 10.104.7.15/16 instead of 10.106.1.11/16?
At this point I would scrub the information from the iSCSI dialogs in ESXi and then uncheck iSCSI to disable it. Reboot the ESXi host and start again. This time make your setup as simple as you can. No authentication. Check your naming etc.
If you haven't already done so I would check in the HP forums to see what you can find.
I agree. Your target IP is stuck in the config. You need to disable iSCSI completely and start over.
-KjB
VMware vExpert
Ok, I'm going to go one step further and just do a reinstall of ESXI. I've already disabled/reenabled iSCSI a couple of times. I'll hope for better luck the second go around. Thanks for the advice.
I'm not sure a reinstall is necessary. Enable and disable isn't the same as remove the information and reboot. iSCSI changes require a reboot. Unless your install went badly I would stick it out and find out where the problem is. iSCSI is well used in ESX(i) so it isn't necessarily something wrong with the install. Anything you can do to further your understanding here will server you down the road. Like when you have a problem and you are in full production.
Ok, I will not go to the extreme of a reinstall yet.
I have cleared out the targets under Dynamic Discovery
Unchecked enable iSCSI
Rebooted Host
When it comes back up, I will reenable iSCSI and configure Dynamic Discovery. At that point, will I need to reboot again before Rescanning?
Just changes require a restart.
Since you have the capability have you considered NFS for storage?
I had considered NFS, but I thought iSCSI was superior? I was going to look at the NFS route if I could not get iSCSI to work.
Bummer, it is still trying to connect to iSCSI via 10.106.1.11, the wrong address.