Hello yaroslav,

for the CentOS7 Starwind VSAN Appliance i can`t choose NVMe as Disk Controller.
I selected now an VMWare Paravirtual instead.

Best Regards

I also think that have to create an systemd service for startup via systemctl at boot.
Like the scheduled task in Task Planer of WIndows Server when StarwindVSA Service is started?

Glad to hear from you again :) To make sure I am reading it right: no rescan happens when you reboot both servers at the same time? Just let me ask questions to let me better understand the situation. Is it a standalone deployment of StarWind VSAN? If it is HA deployment, does the rescan happen loc...

yaroslav (staff) wrote:Hi,

Please make sure to use NVMe controllers and apply this fix https://knowledgebase.starwindsoftware. ... ontroller/

Hello yaroslav,

thank you for your advice, i will implement the change on next downtime.

Good evening, i have setup the rescan script under /opt/StarWind/StarWindVSA/drive_c/StarWind/hba_rescan.ps1 The manual test with sudo /opt/StarWind/StarWindVSA/drive_c/StarWind/hba_rescan.ps1 runs fine. The ESXi host track the login and do the changes to Multipath Policy. However when i restart the...

A short Update on this thread, since i serve the storage to vsan appliance as vmdk no more crashes happened.

Ok, since i cannot believe that an 3 DWPD NVMe is defective after about 5% Usage, i recreated the HA Node this time based on a eager zerod VMDK.
Now the access to NVMe is done from ESXi Kernel.
I will give an update how its going on.

Hello Community, i search for help or self made experience on a eventually Hardware or Kernel Problem. I have two Proliant DL360 G9 installed with VSphere ESXi 7.0.3 Both ESXi running 1 Node of an VSAN Free HA Cluster. NVMe are mapped via passthrough directly to the VSAN VMs. Everything was running ...