Peripherique mort / Chemin mort - Dead Device / Dead Path

[Stephen34]

Bonjour,
J'ai supprimé un LUN dans ma baie NETAPP qui n'était plus utilisé.
De ce fait, celui-ci ne se présentait plus à VMWARE, ce qui est normal.
Mais j'ai un hôte de VMWARE qui voit toujours celui-ci dans Adaptateurs de stockage / Périphériques & Chemins en tant que "mort ou erreur".
J'aimerais donc enlever celui-ci qui ne sera plus jamais contacté.
L'analyse ne donne rien.
Comment je peux faire s'il vous plaît ?

Merci beaucoup par avance,

Cordialement,
-
Hello,
I deleted a LUN in my NETAPP array that was no longer in use. As a result, it no longer presented itself to VMWARE, which is normal. But I have a host from VMWARE that still sees this one in Storage Adapters/Devices & Paths as "dead or error".
I would therefore like to remove this one which will never be contacted again.
The scan yields nothing.

How can I do please?

Thank you very much in advance,
Cordially,

[Stephen34]

I try esxcfg-rescan -d vmhba64:

"Rescan complete, however some dead paths were not removed because they were in use by the system. Please use the 'storage core device world list' command to see the VMkernel worlds still using these paths."

[e_espinel]

Hello.
Try to close the browser and log in again, if not restart the ESXi host.

[Stephen34]

Hello,
It does not matter. I only have the reboot to do...? No way to kill dead things? Restarting is not likely to turn in a loop in order to find the device in question and therefore to put me the ignition of the esx in check? Thanks

[Stephen34]

If I really have to restart my hosts to remove dead devices, I'll have to see this in November at least I think because right now I can't afford to turn off VMs... Anyway, everything works, so it doesn't really matter if I restart later, right?

[e_espinel]

Hello.
Yes, you can reboot when you can, this problem may be due to hardware inventory refresh failure on the ESXi host. Usually this is solved by applying a higher level of patches, but this also requires shutting down the VMs and rebooting.

[pmichelli]

It is likely because of the way you did things. This is the correct way 

1. Evacuate the LUN. Move all VMs off of it and copy any ISO or other files that may be there elsewhere
2. Find the VMFS volume, right click and select Unmount. Check the option to unmount it from ALL hosts
3. Go to storage devices from one host, find the LUN (look for the name that matches the VMFS volume). Select the LUN and click Detach. Check off the box to detach it from all hosts
4. It will now show disabled or look in italics on all hosts. You know it worked if the only option when you click it shows ATTACH
5. Once you have unmounted and detached, at this point you can safely remove the LUN presentation from the NAS / SAN
6. Once this is done rescan the HBA

I've been doing this for over 10 years and never once had an issue

[Stephen34]

Ok thanks, I'll restart when I can, potentially next month or later, I'll have to see...

Good practice has indeed not been respected 😞 I presented it before dismantling it. Something I would pay attention to in the future now! 🙂

Thanks !

[pmichelli]

This guide has pictures but it's essentially everything I posted (yes its old but nothing has changed). Good luck  !

https://vswitchzero.com/2018/10/10/properly-removing-a-lun-datastore-in-vsphere/