ESXi rescan datastore after reboot

[ChrisJ615]

Hi guys,

I hope this is an easy one and that I'm just over looking something.  Whenever I reboot any of my ESXi 5.1 hosts it doesn't rescan for datastores by itself.  It would be really nice for it to do this automatically.  Any advice, help, or direction would be much appreciated.

[a_nut_in]

Hey Chris,

ESX/ESXi actually does a scan and discovery for datastores and LUN's during boot.

This can be seen in the boot.gz file that is populated in /var/log of the ESXi 5.x hosts. So it does not need to do another rescan immediately after boot!

Are you seeing any issues with the LUN's not detected on first boot? Also what version of ESX/ESXi are you running?

Thanks

a

[ChrisJ615]

ESXi 5.1 - and these are iSCSI LUNS.  The interesting thing is I can use the UI or Powershell to do a rescan and everything works.

I am using a custom image from Dell though.

[a_p_]

Just a thought: Do you have spanning tree enabled on the physical switch's iSCSI ports?

André

[ChrisJ615]

Yes, spanning tree is enabled.  But here is the deal: I have two 10G ports that I run VDS's off of over a Cisco 1000v.  (Every host that I have is set up that way.)

[Josh26]

You should still be able to configure portfast or "edge" depending on your physical switch. The 1000v doesn't send BPDUs and won't trip anything. STP negotiation time is probably your issue.

[ChrisJ615]

Hey Josh,

Let's say that STP negotiation time is my problem.  How could I trouble shoot it?  I guess that is what I'm looking for.  I know what symptoms I'm having.  Now I need to be able to troubleshoot the root cause.

[Josh26]

The only real troubleshoot here would be to portfast the port (best practice: it should have been done anyway) and see if it fixes the problem.

[mcblind]

we had the same problem on ESXI 6.0. we just the steps below and it worked.

1. I manually added the datastores again, then removed the VMs from inventory.
2. Reboot.
3. Re-add datastores, specifying "Assign new signature"  Takes a few seconds for each.
4. Re-add the VMs using Browse datastore, and adding the .vmx
5. Reboot.  Datastores are there.  YES.
6. One VM wouldn't start because it had a second hard disk which was on the second datastore.  Edit the VM, remove the second HDD, re-add it, specify an existing HDD.
7. When starting the VMs, I was asked if I copied or moved them.  Selected "Moved".
8. Added VMs to Automatic Startup, rebooted once more for good measure.  All good.

[mbfromit_1]

We had a similar issue.  ESXi 5.5.   Our revolved around boot times ( or at least it appears )
We have a SAN that takes about 10 mins to boot.  The 3 ESXi Hosts ( Dell's ) take only 3-5 mins to boot

So if the power goes out and the outage exceeds our UPS capacity and thus the systems all shut down,

when the power does come back up, all 3 Hosts are up and running prior to the SAN becoming available

Anyway to affect the boot order of things (Dell servers after EMC SAN) in hopes that the SAN boots first

so it's available when ESXi 5.5 is going through its "first scan" during its boot process?

[jpf1]

mbfromit_1 Did you ever get a solution for this?