Backup target - 2 or 4 drive NAS?

gjacobse

When I was replacing the NAS (single drive, proprietary format) I looked at some of @scottalanmiller's comments on how much was my data worth.

I went with a ReadyNAS 4 drive system running RAID10.

I may only be using it to start photos, some music and movies,.. but some of it I can't replace. Can't find or get to photos taken in the late 1800's or early 1900's from family in Europe.

I still need to do a off-site backup,.. but I have thought of a few ways to do so...

Dashrender

@scottalanmiller said in Backup target - 2 or 4 drive NAS?:

@Dashrender said in Backup target - 2 or 4 drive NAS?:

Read/Write - 112/112

Write is half the speed of reads, not the same.

This is a claim from Synology's site.
https://i.imgur.com/EiXISx6.png

scottalanmiller

@Dashrender said in Backup target - 2 or 4 drive NAS?:

@scottalanmiller said in Backup target - 2 or 4 drive NAS?:

@Dashrender said in Backup target - 2 or 4 drive NAS?:

Read/Write - 112/112

Write is half the speed of reads, not the same.

This is a claim from Synology's site.
https://i.imgur.com/EiXISx6.png

You need to read these things more closely. Look again and see if you can spot where you are totally wrong and totally misunderstood their marketing.

Dashrender

@scottalanmiller said in Backup target - 2 or 4 drive NAS?:

Two drive RAID 1 setup is cheaper and safer. Uses less power (about 11W less), generates less heat (tiny) and takes up less space (tiny difference.) But other than IOPS, it wins in every way.

Sure, I understand this in theory, but haven't figured it out in practice.

My current backups go to a Gen 1 Drobo 8 Bay SAN (iSCSI) that has to be shared off a server as Drobo doesn't support more than one iSCSI connection. Plus it's running Beyond RAID on 5 drives with a one drive fail allowance... It's pretty darn slow - probably slower than a single drive on it's own.

scottalanmiller

I'm writing a response but waiting for you to figure out all of the things that you misunderstood.

scottalanmiller

@Dashrender said in Backup target - 2 or 4 drive NAS?:

Sure, I understand this in theory, but haven't figured it out in practice.

What do you mean? The smaller a RAID 0 is, the safer it is. It's that simple.

scottalanmiller

@Dashrender said in Backup target - 2 or 4 drive NAS?:

My current backups go to a Gen 1 Drobo 8 Bay SAN (iSCSI) that has to be shared off a server as Drobo doesn't support more than one iSCSI connection. Plus it's running Beyond RAID on 5 drives with a one drive fail allowance... It's pretty darn slow - probably slower than a single drive on it's own.

HOw does this relate to the associated statement? I'm unsure what the example is supposed to be telling us.

Dashrender

@scottalanmiller said in Backup target - 2 or 4 drive NAS?:

@Dashrender said in Backup target - 2 or 4 drive NAS?:

@scottalanmiller said in Backup target - 2 or 4 drive NAS?:

@Dashrender said in Backup target - 2 or 4 drive NAS?:

Read/Write - 112/112

Write is half the speed of reads, not the same.

This is a claim from Synology's site.

You need to read these things more closely. Look again and see if you can spot where you are totally wrong and totally misunderstood their marketing.

Average? ug.. ok - another meaningless thing.. thanks.

Dashrender

@scottalanmiller said in Backup target - 2 or 4 drive NAS?:

@Dashrender said in Backup target - 2 or 4 drive NAS?:

Sure, I understand this in theory, but haven't figured it out in practice.

What do you mean? The smaller a RAID 0 is, the safer it is. It's that simple.

What I was saying/meaning is that I haven't done an IOPs study to see what anything I'm doing it using IOPs wise. i.e. I've never run DPACK to see what my server reports it's IOPs to be doing, etc.

scottalanmiller

@Dashrender said in Backup target - 2 or 4 drive NAS?:

@scottalanmiller said in Backup target - 2 or 4 drive NAS?:

@Dashrender said in Backup target - 2 or 4 drive NAS?:

Read/Write - 112/112

Write is half the speed of reads, not the same.

This is a claim from Synology's site.
https://i.imgur.com/EiXISx6.png

So here is what I spot quickly:

I was wrong here, Synology makes the claim, but it's clearly nonsensical and must be a typo. It's impossible for that device to do that and impossible for them to know what you will get with components that they know nothing about.
Read the metric of the chart, that's not a drive speed metric, that's a networking metric.
You are seeing the speed of the NIC on the unit, not the drives. That chart is unrelated to the disks, RAID or storage. It's simply the network speed of the device.
You quoted RAID speeds but Synology isn't and can't tell you those. Those are determined by the specific drives in question and the RAID level. Synology can't know those things unless the chart is about your drives in particular. Synology doesn't make any claim that I see to telling us about RAID or drive speeds.
Storage performance is in IOPS and is a factor of the drives that you purchase far, far more than the NAS. The NAS has limits, but those should be worlds beyond the potential speeds of the drives. Millions of IOPS instead of hundreds or scores.
Mirrored RAID is always 1:2 no matter how you slice or dice it. This is something you just know, not something you ask a vendor about. Even looking at Synology for this info was fundamentally wrong.

scottalanmiller

What Synology has done, to make this claim kinda legit, is look at what disks "can" stream (which is more than is listed here) and added the "cap" of the network. So if you do a contrived operation that pushes the drives to their throughput limit (a useless number hence why we don't measure drives by that metric) but tells us nothing about performance. That could be just two or three IOPS producing that limit. But in the real world, that's not useful.