I've been facing such a weird problem. I'm having hard time in creating a VM on a shared datastore in vCenter Server 6.5. I would like to provide some information about my environment which I've set up in my Lab --
Here's my Problem:
When I create a VM (irrespective of ESXi Server) on shared either datastore iSCSI-LUN-0 and iSCSI-LUN-1, it fails every time with an error message :
" A general system occurred: pbm error occurred during precreatecheckcallback: unexpected soap fault: ns0:requestfailed; request failed"
Note: This only happens when I create a VM on shared datastore. I face no problem while creating VM locally on the ESXi Host itself.
Please lemme know if any further information is required to drill down.
Any help will be greatly appreciated.!
Do you have any local drives on ESXi that you can try to build the VM on? This will rule out the Starwind virtual SAN being the problem.
I do have 100 GB Local Storage for each ESXi Server. But I've to enable DRS and HA Cluster in my Lab for testing. DRS and HA do not function if VM resides on local storage of ESXi Server.
Yes but as a test are you able to deploy a VM to that storage?
Yes, I'm able to deploy a VM on local storage of ESXi Host. But when I try to deploy a VM on shared datastore, I get an error message.
In order to test DRS and HA, VM has to be deployed on shared datastore.
I've to do some testing on DRS and HA. That can only happen when I can deploy a VM on shared datastore.
Usually, problem happens on nework when you use iSCSI storage.
Can you check and post vmkernel log collected when you get this error.
Also Can you check that MTU is correctly set (the same size) on all points between ESXi host and storage array.
I would recommend changing TCP parameters on StarWind server, this should solve the case.
Additional info you could find here. https://forums.starwindsoftware.com/viewtopic.php?f=5&t=2293
Also, from my experience, that's a good idea to check the MTU on all instances, even on switches if you are using them between storage hosts.