Dear team,
Request you to help me with the following queries?
Q1 :- Is UUID and Device Id is same?
Q2 :- IS device Id provided by storage array to a LUN?
Q3 :- Is UUID provided by an ESX server so that ESX server can identify that LUN?
Q4:- Is UUID also called or refers ESX signature?
Regards
Mr VMware.
Hi,
in general, the signature shouldn't change in a scenario you desrcibed.
The ESX host used to create a VMFS datastore on that LUN does collect information provided by the array.
These information will be used to generate a signature (the process itself is still not published by VMware).
The signature itself is stored inside the VMFS datastore and will be always used when performing SCSI rescan operations.
Whenever an ESX Server perform a reboot or rescan operation, the following procedure will be used (high level).
The ESX Server will
So you could easily allow other ESX Servers access to this device when each ESX Server does see the device with the same SCSI information.
As a best practice, each ESX Server shoud see such shared device with the same LUN ID.
Modern arrays are capable to present storage devices with different LUN ID's on an intiator (HBA) base.
Another thing to keep in mind that the used array front end ports are using the same SCSI Port/Flag Settings, but these flags might differ between vendors.
If these two basic requirements are met, you will be able to share the storage device easily between multiple ESX Servers.
Kind regards,
Ralf
Need assitance from VMware Experts...
Hi,
if your question belongs to storage devices, than the answers are.
Q1: no
Q2: yes
Q3: yes
Q4: unable to answer this question because unaware what you would mean with ESX Signature
Check the following article which do provide detailed background information.
Identifying disks when working with VMware ESX
Kind regards,
Ralf
Ralf - he means the VMFS signature (for Q4), and the answer is sort of yes. The signature is written to the volume, and *part* of that signature is the UUID.
The rest of what ralf said is spot on.
can u plz help me , Signature includes what information?
if a LUN is need to be shared to second esx host , will 2nd esx host overwrite the 1st ESX host signature???
Hi,
in general, the signature shouldn't change in a scenario you desrcibed.
The ESX host used to create a VMFS datastore on that LUN does collect information provided by the array.
These information will be used to generate a signature (the process itself is still not published by VMware).
The signature itself is stored inside the VMFS datastore and will be always used when performing SCSI rescan operations.
Whenever an ESX Server perform a reboot or rescan operation, the following procedure will be used (high level).
The ESX Server will
So you could easily allow other ESX Servers access to this device when each ESX Server does see the device with the same SCSI information.
As a best practice, each ESX Server shoud see such shared device with the same LUN ID.
Modern arrays are capable to present storage devices with different LUN ID's on an intiator (HBA) base.
Another thing to keep in mind that the used array front end ports are using the same SCSI Port/Flag Settings, but these flags might differ between vendors.
If these two basic requirements are met, you will be able to share the storage device easily between multiple ESX Servers.
Kind regards,
Ralf
DEAR RALF SIR,
many thanks for the detailed and clear explanation.
Sir last question? whenever we request to map an exiting lun to a 2nd ESX host, once we rescan the adapter it' won't mount automatically becasue of mismatch LUN id (LUN detected as a Snapshot LUN) what mistakes this SAN people were doing while mapping a exisitng lun to a 2nd ESX Host.so that we can avoid such issues in future.....
Regards
Mr Vmware
Hi,
as I'm not familiar with IBM storage I'm not really able to answer this question.
IMHO, this could be caused by
See followig VMware article which does provide detailed background information.
vSphere handling of LUNs detected as snapshot
But to be honest, normally ths is a pure storage admin task and can't be solved from the VMware side.
So you need to address this to the storage team, even if they're still not up to speed.
Good luck,
Ralf