How to change / replace a failed “system” disk on EMC Clariion CX500

mrdenny's response is the right one. Ideally the Navisphere Host Agent will do this for you but if that is not working (give it a couple of minutes and a restart if possible) or not an option you can do this solely within Navisphere's Connectivity Status dialog.

Within Navisphere right click on the Array and select Connectivity Status. If your host is powered on and your Fabric Zoning has been correctly installed you should see entries with the WWN's (the hexadecimal strings you mention) of the HBA's from your server listed as logged in but not registered. Select these and hit the Register button. At this point you can select an existing host if you've already set one up or you can create a new host identity to map the WWN to. If you have a standard redundant setup with two HBA's routing through two separate Fabrics that are in turn connected to separate front end ports on each of the Storage Processors - this means you will typically have 4 entries to manually register. If the host was already registered with different HBA's you should remove these at the same time (assuming they have now been completely removed) in order to keep things neat.

Once you have the HBA's registered (each entry shows as logged in and registered in the Connectivity Status dialog) you can then check that the correct LUNS are assigned to the Storage Group for the host. If you are editing an existing Storage Group then you may need to refresh that directly by checking the Connectivity Status dialog for the Storage Group itself.

At this stage any LUNS in the storage group should be visible to the server - rescan for storage within the OS and it should be visible. Some OS's need a couple of scans for this to work - ESX is particularly sluggish if you rescan from within the VI Client for example.

You should be able to do this the same way you can fix problems with multi-homed servers where the agent mistakenly registers an ip-address that is inaccessible to the array's management interface (e.g. where a private cluster\heartbeat network address gets registered rather than the correct management interface address). For systems that are automatically registered using the Navisphere Agent the best way to do this is by updating the AgentID.txt file on the Servers and restarting the Navisphere agent as per these instructions.

For systems where you are not using an agent then using Navisphere Manager to deregister the old hosts addresses and recreate them with the new addresses - you will then need to ensure that the Storage Groups they belong to are still OK as they you will have to reassociate the servers with the appropriate Storage Groups again.