These are existing hosts that can see luns that were already there prior to the creation of these new luns.

this is an existing host that can see LUNs and you are trying to add another LUN or this is a brand new host ?

I have a storage group for each server.

How do you tell a lun's HLU number? Is it identified some other way in Navisphere? It's very possible that one of your senarios regarding HLU is plausible.

We don't have 255 luns, so I don't believe that is the case.

Did you add the LUNs to the Storage Group that the Hosts are in?

Do you have one Storage Group for the whole ESX cluster or a Storage Group for each server?

If you have multiple storage groups, you need to make sure that each array LUN has the same HLU# in every storage group or ESX will not be able to identify the LUN correctly in the cluster.

After adding the LUNs to the Storage Group, you need to Rescan disks in ESX to see the new LUNs, then create new datastores on those LUNs.

If you the LUNs don’t show up, it’s possible that the maximum LUN count in ESX has been hit (there is a hard limit of 255 but there is also an advanced setting that sometimes has a lower value.)

Another issue could be if your first set of LUNs have HLUs of say 0 through 10, and then you add a new set of LUNs from 20 to 30, but there are no LUNs with the ID’s of 11 to 19, ESX used to have a default setting “sparseluns = 0” that it would stop scanning when it stopped finding LUNs so it would never check for the ID’s above 11 (in this example).

Without a lot more configuration information, it would be hard to really know what is going on.

Richard J Anderson

in storage group properties, LUNs tab .. on the right hand side there is a column (Host ID).

Sorry for the lack of detail.

Yes. I do a rescan on the host.

In the past, all I've had to do is put the luns I want available to the ESX host into a storage group in Navisphere. I would then attach the ESX host in Navisphere to the storage group I created then do a rescan on the host in vSphere and the luns would show up attached to the host.

For some reason that doesn't seem to work this time

Am I missing a step. Or is this not the correct way to go about this?

sorry for asking the obvious, you did rescan in ESX server ? After that when you try to craete new datastore you don't see the LUN ?

If multiple hosts are sharing the same LUNs/Datastores, then each array LUN must use the same HLU on every host.

For example,

You have two hosts ESXA and ESXB that are in an ESX Cluster, and each host has a separate storage group in Unisphere, labeled ESXA_SG and ESXB_SG. Then you have LUNs 10 and 11 in Unisphere that you want both servers to use.

When you add LUNs 10 and 11 to ESXA_SG, they get HLUs 0 and 1 (for example). Now when you add the LUNs to ESXB_SG, you must manually set the HLUs to 0 (for LUN10) and 1 (for LUN11) or the ESX servers will get confused.

Richard J Anderson

Sr TC - Western Division, EMC

richardj.anderson@emc.com

206.229.9263

For the sake of trouble shooting let's say I have Storage Group A and Storage Group B.

Both Storage Groups have 20 different luns in them.

When I go to Storage Group A's Properties, LUNs tab and look under the Host ID they are labled 0-19.

Now when I go to Storage Group B's Properties, LUNs tab and look under Host ID, they are also labled 0-19, even though they are different LUNs.

Could this be part of my problem?

is that still a requirement ? I thought with introduction of NAA (Network Address Authority), LUN IDs did not matter anymore ?

It’s possible I suppose, but I hadn’t heard that yet.

Richard J Anderson

Is the ESX Host HBA zoned to both SPs? It should be…

If the zoning is correct, then the host connections on that sP may not be enabled in connectivity status in Unisphere. You may also need to go into Engineering mode to add those paths to the Storage Group if you find that not all of the host paths were already added.

Richard J Anderson

It looks like the luns I can't get to work are all on one SP. I didn't set them up this way origianally. The luns I originally put on the other SP show that SP as their default, but they are no longer on it.

Could it be that this SP has failed?

I did an SP collect, but didn't see any indication that the SP failed, nor is there an "F" above that SP.

It seems wierd that all working luns are on one SP, and all luns I can't get to work are on another.

As a test, you can right click on the LUNs that you can't see from the host and select Trespass. That will move the LUNs to the other SP that contains the LUNs that the host can see. If after trepassing the LUNs you can now see them on the host, then you have an issue with the host not able to see the other SP. If that is the case, in the Storage Group for the host, expand the tree view, and in the Host section, you should see your host listed, right click on the host and select Connectivity Status. There you should see the host and the paths - all paths should be listed in the Storage Group name - if they are not, then click on "Reconnect". That should attach the paths to the storage group.

glen

how big is that LUN ?