[pve-devel] scsi-hd vs scsi-generic with iSCSI

[Thomas Lamprecht]

On 10/7/19 5:45 PM, Daniel Berteaud wrote:
> 
> 
> ----- Le 2 Oct 19, à 18:41, Daniel Berteaud daniel at firewall-services.com a écrit :
> 
>> ----- Le 30 Sep 19, à 11:52, Thomas Lamprecht t.lamprecht at proxmox.com a écrit :
>>
>>>
>>> Depends on the outcome of above, but effectively we would like to
>>> not have the choice between "working or not", so ideally we can
>>> just either fix it in QEMU or set the respective (working) option
>>> for all new started VMs (QEMU machine versioned, to not break live
>>> migration due to changed VM HW).
>>
>> Just thinking about it, as scsi-hd, scsi-block scsi-generic won't present the
>> same HW to the VM, moving disk from one storage to another is already broken if
>> both source and dest are not the same type (eg if source uses scsi-hd but dest
>> scsi-block or scsi-generic). Without even considering the issues I have when
>> ZFS is used as a backend.
> 
> 
> Just opened a new bug : https://bugzilla.proxmox.com/show_bug.cgi?id=2408
> Moving disk between storages which would use different backend (scsi-hd, scsi-block or scsi-generic) make next migration crashes the guest. Not directly linked to ZFS over iSCSI, it's a more general issue.
> 

Thanks for the nice write up and clear reproducer!

It seems that if we cannot use the same backend for all disks we need to
die when a disk move to a storage backend is request, and that move would
need to change the scsi "backend".
As I'd not like to die it would be better to see if there's still the need
for different backends.

cheers,
Thomas