LUN scan doesnt show new luns until reboot?

[Yattong]

hi

quick background, 16 esx 3.0.2 with eva 6000.

Whenever we create a new lun, or re-size a lun or what ever... we go through the correct procedures, i.e removing from datacentre / hosts and rescanning the lun.

The question is... sometimes the new LUN does not appear, no matter how many times I try to scan it, either through the VC or by command prompt. (esxcfg-rescan). It does appear however, after a reboot. No scanning needed after.

Is there something I can restart (not server reboot), so I can then do another scan for it to appear?

Ive tried rebooting the mgmt-vmware. Will try vxpa next.

[vlchild]

Had this problem too. FastT backend, IBM servers with QLogic HBAs.

This is how we worked around it:

Go to the first host.

Configuration Tab

-> Storage Adapters

Right click on one of the HBAs

Click 'Rescan'

The LUN should show up.

(Side note: clicking 'rescan' at the top to rescan all HBAs at once does NOT work)

One get it formattted for VMFS you will have to do that for every host. ALSO, I had to do this to get the multipathing setup correctly. The dual paths did not show up in the 'Storage' area either.

Lee

[bobross]

We had this problem once also...on an EVA (which we have since gotten rid of)...notice your mpath.txt file has out-of-order LUNs. It skipped over LUN 39, scanned to 42 and then discovered 39. Once you get this out-of-order LUN, the scan will stop, so you will never see LUN 43. It's a timing issue. Other systems have the concept of 'persistent bind' to alleviate the issue. You may get lucky and find all the LUNs in order, but most of the time you won't. You can also try the HBA rescan (as the previous poster mentioned) to see if it helps. The chances of getting an out-of-order LUN during the scan increase with the # of LUNs.

[mikepodoherty]

What I have seen is that with either the IBM DS4000 series or the EMC CLARIION or DMX SANS - fibre channel connections - if you rescan on the host, VirtualCenter doesn't see the new LUNS and I have to rescan in VirtualCenter.

I've tried restarting vpxa andmgmt-vmware - it is simpler to have VC open, rescan on the ESX server and then follow up with a recan in VC.