vSAN Free @ Proxmox - The replication partner is not synchronized

[yaroslav (staff)]

ZFS for production is not good due to a huge performance hit. HDDs make it even worse.
Try creating a physical array out of those 6 disks. Here, some remote assistance will be great to review and help you with reconfiguration, yet it is not possible as you have a FREE license. Please discuss converting to StarWind VSAN paid with the person who provided you with the key.

Where I can & download the Starwind Windows console, because it wasn't bundled with the Proxmox package?

Press the gear-shaped icon > navigate through the menu to Install Management Console > you should be able to download the installer. Make sure to select only the StarWind Management Console and Integration components library when installing.

[dwma]

ZFS for production is not good due to a huge performance hit. HDDs make it even worse.
Try creating a physical array out of those 6 disks. Here, some remote assistance will be great to review and help you with reconfiguration, yet it is not possible as you have a FREE license. Please discuss converting to StarWind VSAN paid with the person who provided you with the key.

Where I can & download the Starwind Windows console, because it wasn't bundled with the Proxmox package?

Press the gear-shaped icon > navigate through the menu to Install Management Console > you should be able to download the installer. Make sure to select only the StarWind Management Console and Integration components library when installing.

Ok, I'll sacrifice the the environment, and do the RAID way (2x RAID1 for OS and 6x RAID6 for vSAN).
Could you tell what is the best approach to give the CVM access to the RAID6 datastore?

Passthrough as I've done previously for the disks
qm set 903 --scsi1 /dev/disk/by-id/wwn-0x600508b1001c52ecxxxxxxxx

or add LVM on top of the RAID6 array, and create one big qcow2 disk image for CVM?

[yaroslav (staff)]

No need to sacrifice the environment if the reconfiguration is done to one node at a time, starting with the not synchronized one.
RAID0 seems a bad idea given no redundancy. Go for RAID1.
RAID6 out of HDDs is against best practices (https://www.starwindsoftware.com/best-p ... practices/). If you can't do RAID10, try RAID5 so the write penalty is lower. Again, parity RAIDs out of HDDs are agains best practices.

qm set 903 --scsi1 /dev/disk/by-id/wwn-0x600508b1001c52ecxxxxxxxx

This one seems to be the best option for the physical RAID to be passed through.
As a side note, I think you are confusing pass-through with RDM. What you did was connect devices as RDM. I want you to do that to the physical RAID again. AFAIK, you can't pass through the individual disks as they are backed behind the RAID array. I might be wrong here, but pass-through in terms of storage is configured differently and requires reconfiguration in Proxmox grub config.

[dwma]

No need to sacrifice the environment if the reconfiguration is done to one node at a time, starting with the not synchronized one.
RAID0 seems a bad idea given no redundancy. Go for RAID1.
RAID6 out of HDDs is against best practices (https://www.starwindsoftware.com/best-p ... practices/). If you can't do RAID10, try RAID5 so the write penalty is lower. Again, parity RAIDs out of HDDs are agains best practices.

qm set 903 --scsi1 /dev/disk/by-id/wwn-0x600508b1001c52ecxxxxxxxx

This one seems to be the best option for the physical RAID to be passed through.
As a side note, I think you are confusing pass-through with RDM. What you did was connect devices as RDM. I want you to do that to the physical RAID again. AFAIK, you can't pass through the individual disks as they are backed behind the RAID array. I might be wrong here, but pass-through in terms of storage is configured differently and requires reconfiguration in Proxmox grub config.

I've remade storage on all proxmox hosts, went back from HBA mode into raid mode (RAID1 2x disks for Proxmox and 6x RAID10 for Starwind vSAN - these are RDM on each starwind vsan vm). Configured networking, added all 3 appliances inside Starwind panel, but I'm failing with one host on the Storage pool (2 hosts sucessfully created its own storage pool, one host cannot, even after restart). I've uploaded the logs from the faulty host into the case #1364582

[yaroslav (staff)]

Can you please spare the error you are getting so I can look into the logs?
Just to make sure that I read it right. You connected RAID into the VM as physical device, and you are attempting to format the storage now (i.e., create a pool).
Does the disk show as online or ready in the disks?

[dwma]

Strange... today morning it created the pool without any issue...

Yes, exactly - I've made a "physical" RAID (on P440ar raid controller) and it's connected to the VM as physical device.
On StarWind vSAN side - the disk was in Ready state, but previously when creating a pool on one host it failed with a generic error like create pool failed.

[yaroslav (staff)]

Strange. It could be something like background initialization or so that prevents discard from running well.
That's just my humble opinion. Glad to read it works now tho.