Adding SG//LUN to RHEV

Hi,

Have you de-commissioned the ESXi and then installed RHEL on the same server ? or ESXi is already existing and he is trying to add another RHEL server having the same host name?

Thanks

Jyothi

If you're reusing the same IP addresses, try shutting down the hosts, then remove the hosts from the Storage Group, then de-register the hosts, then restart the Management Server on both SP's from array_IP_SPx/setup (restarting Management Server will force the array to look reset/rescan the hosts). If you use the same IP as the ESXi hosts the registration may be persisted on the array. If you're not using the EMC Host Agent for Linux then you'll have to manually register the hosts again (I recommend you use the EMC Host Agent).

This may not work because when an iSCSI host is powered down, the registration disappears from the Unisphere GUI.

glen

Hi thanks for the input...  you are right, these were esxi hosts and now rhev...  I did deregister them and got them to go away completely in the EMC console.  I then manually added them again... but I didn't shut them down and I didn't restart the SPs.

Unfortunately, as I am using the Hypervisor version of RHEV which is essentially stateless, I can't load the emc host agent - or I don't think I  can anyway, it will disappear after a reboot...  but I need to use this RHEV-H as I call it, because if I use the version that runs on RHEL, then I have to pay for RHEL license as well as RHEV and then it's nearly as expensive as VMware...

I have a few things to look at yet... do I need to trespass everything from one SP to the other if I am restarting it? or does it do that itself?  I'm not exactly an expert with the VNX, I know how to do specific things very well that I have learned to do over the years.  But I am not a storage expert.

Thanks again

Bill 

Bill,

Are you actually using storage from this VNX ?  If you are presenting storage to this or other system you need to make sure that host multi-pathing software is configured (dm-mpio, powerpath ..etc) before you reboot SPA/SPB. When you reboot an SP all LUNs will trespass to its partner.

I recommend restarting the Management Server - this is a restart of the GUI Unisphere, not an actual reboot of the SP's. This does not affect the LUNs or the hosts connectivity. To access the Restart function, go to https://IP_address_SPa/setup. When the login screen comes up log in using your Unisphere login and once you're on the setup screen, scroll down a bit and you'll find a "Restart Manager Server" button. Clicking this will restart the software on the array, but will not affect the operation of the LUNs. It takes about 60 seconds to come back up (you can try to open Unisphere using the same IP address for SPA to see when it is up), then do the same for SPB and wait for it to come up also. Then check on the Host/Initiators to see what is there.

glen

hey all,  I provided VPLEX storage devices to a RHEL/Linux customer hosts on a Dell blade, LUNs starting at 0 - 199.  Customer reports:

RHEV Hypervisor has trouble recognizing a LUN number as 0 and wants me to pull the dev LUN 0 and renumber it.  I'd rather not renumber our storage as we provision to hosts, not to Hypervisors.   Anybody have any thoughts on the Hypervisor issue? It's RHEL x86_64 7.2.

Thanks