ReFS on iSCSI LUN hosted on a NAS

[b.vanhaastrecht]

Hi,

Is a ReFS volume on an iSCSI LUN hosted from a Synology NAS a good thing to do?

The reason for this is to gain performance by utilizing faster synthetic fulls.

It's a fairly strong NAS with four 2.1Ghz atom cores and 12x8TB drives.

Kind regards,
Bastiaan

[PetrM]

Hi Bastiaan,

I don't see a problem with that as long as it is supported by Synology, at least it's recommended to confirm details with the storage vendor, here is an article from Microsoft.

According to this post, the configuration with iSCSI LUN is stable enough but if I'm not mistaken there was an issue with possible data corruption when Write Cache was enabled. There is an example on Synology forums, I believe your question worth asking there as well.

Thanks!

[Gostev]

It is certainly a way better than using SMB, from every perspective but still not as ideal comparing to using a general-purpose server with internal or direct attached storage. I covered this in details in one of my newsletters last fall, but here's a shorter summary of that story that I had to make for my Reddit post on the same exact subject a few days ago:

Yes, moving to iSCSI completely removes the chance of data corruption in the SMB stack specifically.

You can then further improve reliability by enabling iSCSI header and data digests in the SAN settings. This will address the majority of data corruption/loss events which may happen to data in transit due to faulty network equipment, which are fairly common too. This however still leaves the possibility for corruptions to occur during the protocol crossover between Ethernet, TCP and iSCSI.

This is why it is so important to have integrity checking as a part of a higher-level protocol, as is the case with Veeam's proprietary data movement protocol. Which in turn makes general-purpose servers with internal storage the best possible candidate for a backup repository in terms of reliability. Because as soon as data is stored on some other box, it has to travel there over the network, at which point Veeam loses control over it and is no longer able to ensure its integrity.

[karsten123]

hi all,

i have a small standalone Hyper-V environment and a NAS (sorry. not my design. i am angry, too) for the new repository. the vbr server is remote and managed by us.

my idea is to attach the NAS via iSCSI to the Hyper-V host and format it with ReFS.
the main goal is to avoid SMB for sure and get block clone capabilities.

what are your thoughts?

Tia
Karsten

[david.domask]

Hi Karsten,

Sure, this can work. I've merged your post with an existing topic on the subject, and the resources posted above should assist