VMware ESXi very slow boot with RDM LUNs used by MSCS
When RDM (Raw Device Mappings) are used in in VMware vSphere, my 2 ESXi host (VMware ESXi 6.0.0 4192238) can take a very long time to boot (30 minutes), when i rescaned datastores ESXi server freezed. I tryed disconnect FC cables and iSCSI cables, same problem.
- updated 07/10/2016 – it’s doesn’t matter if you have FC or iSCSI connectivity, problem is with both
Solution is very simple, first of all you need enable SSH to the affected ESXi host, you can enable it in console (troubleshooting options) or via vSphere Client/vSphere Web Client in Configuration of ESXi host >Security Profile > Properties > Start SSH Service. How to Enable SSH on ESXi host.
esxcli storage core path list
Run Putty or your favourite SSH client and connect to affected ESXi host. When you are logged in, you need list LUNs (device), now identify naa.xxx of your RDM mappings.
fc.20000024ff7c1f4a:21000024ff7c1f4a-fc.500000e0da097040:500000e0da097021-naa.600000e00d2800000028097000040000 UID: fc.20000024ff7c1f4a:21000024ff7c1f4a-fc.500000e0da097040:500000e0da097021-naa.600000e00d2800000028097000040000 Runtime Name: vmhba2:C0:T0:L4 Device: naa.600000e00d2800000028097000040000 Device Display Name: FUJITSU Fibre Channel Disk (naa.600000e00d2800000028097000040000) Adapter: vmhba2 Channel: 0 Target: 0 LUN: 4 Plugin: NMP State: active Transport: fc Adapter Identifier: fc.20000024ff7c1f4a:21000024ff7c1f4a Target Identifier: fc.500000e0da097040:500000e0da097021 Adapter Transport Details: WWNN: 20:00:00:24:ff:7c:1f:4a WWPN: 21:00:00:24:ff:7c:1f:4a Target Transport Details: WWNN: 50:00:00:e0:da:09:70:40 WWPN: 50:00:00:e0:da:09:70:21 Maximum IO Size: 33553920
In my case it was 3 devices:
- naa.600000e00d2800000028097000020000
- naa.600000e00d2800000028097000030000
- naa.600000e00d2800000028097000040000
esxcli storage core device setconfig -d naa.600000e00d2800000028097000020000 --perennially-reserved=true esxcli storage core device setconfig -d naa.600000e00d2800000028097000030000 --perennially-reserved=true esxcli storage core device setconfig -d naa.600000e00d2800000028097000040000 --perennially-reserved=true
We need run script, which Enable the Perennially-Reserved Setting for these 3 devices and run it on affected ESXi hosts.
To verify, that startup process of ESXi host has become quicker you need reboot the ESXi host. After reboot try rescan datastores.
Dont forged stop SSH Service on your ESXi hosts.
- Update 07/10/2016
Third server Fujitsu RX200 S8 with latest VMware ESXi 6.0.0 4192238, only 2-port 10 GbE card and software iSCSI adapter. Without perennially-reserverd booting for 4,5 hour… With Perentially-Reservet Settings booting for 10 minutes, so lets try change settings of iSCSI with Static Targets instead of Dynamic Targets. now 6 min to boot.