BUILD My 1st Server Build Ever

[Arman]

RAIDZ3 is slower than RAIDZ2, which is itself slower than RAIDZ1, and uses more of your disk space for parity. In return, it offers greater data protection. It's up to you whether the greater data protection is worth it, but I don't believe it generally is. I believe that most of the time, for the home user, RAIDZ3 is a waste of CPU cycles, disk space, energy, and consequently money. The money would be better spent on (1) a lower-spec machine to act as a backup server, ideally stored offsite; or (2) some form of cloud backup solution.

No, there is no such feature, and it's unlikely iXSystems will be able to add one. This is baked into the core of ZFS.

The screen shots you posted indicate, based on the assumptions you entered (which include an MTBF of 500,000 hours for the drives, and that you'd replace a failed drive within 72 hours), that a RAIDZ3 vdev consisting of five, 2 TB drives will, on average, go 114 billion years (that's what 1 x 10 ^ 15 hours means) before you'll lose data to disk failures.

No, because I'm not able to reach @Bidule0hm's calculator from here. But if you want to look at the numbers for a 6 x 4 TB RAIDZ2 vdev with 500khour MTBF (mine are WD Red drives, so it's actually 1Mhour, but I doubt it will make a difference), you'll see that the MTTDL line is still a very large number.

In My Opinion.

Ohh I understand what you're saying now. Alright, this is my argument:
Everyone has their priorities. Some people value speed more, some people value money and energy more. For me, I value my data the most. That's just my personality. I care a lot about my memories so I would take a few more steps onto the safer side just in case...

The money would be better spent on (1) a lower-spec machine to act as a backup server, ideally stored offsite; or (2) some form of cloud backup solution.

The money saved wouldn't be that significant enough to allow me to purchase an offsite backup system, however it would help a bit.

For now I do not need so much space to be annoyed by the fact that I would get 40% available space with Z3 compared than 60% with Z2... I could just buy bigger capacity in the far future if I need more space. Samsung recently released their 16 TB Flash storage which in the future will force the prices of HDDs to drop significantly. (Which then would allow me to get something like ten 8TB WD Reds for a much cheaper price if I ever needed so much storage)

No, there is no such feature, and it's unlikely iXSystems will be able to add one. This is baked into the core of ZFS.

@Mirfster Couldn't I just rename the files? Wouldn't that cause them to be rearranged onto the new VDEV too? @Mirfster You did say that if the file is "modified" It would...

 

[danb35]

I'm wondering if vdev and zpool are being used for the same thing in some of these posts.

If so, that's being done incorrectly. A pool (or zpool) consists of one or more vdevs, which in turn consist of one or more devices.

Isn't a zdev part of a zpool and as such spanned across all drives in said zpool?

I assume you mean a vdev. No, that's not how it works. A vdev, as I mentioned above, consists of one or more devices. When a pool consists of more than one vdev, data is striped across all vdevs in the pool.

 

[Arman]

I'm wondering if vdev and zpool are being used for the same thing in some of these posts. Isn't a zdev part of a zpool and as such spanned across all drives in said zpool? Or did I miss some context in some earlier posts.
Also bit rot is not much of a factor if you have disk scrubs scheduled at least once a month.
But disc failure in regards to RAIDzX has been spot on in my understanding.


Sent from my iPhone using Tapatalk

Go check this out: Slideshow explaining VDev, zpool, ZIL and L2ARC for noobs!
It should clear up any confusions you have.

 

[depasseg]

If this statement is true:

For me, I value my data the most. That's just my personality. I care a lot about my memories so I would take a few more steps onto the safer side just in case...

then this statement confuses me:

The money saved wouldn't be that significant enough to allow me to purchase an offsite backup system, however it would help a bit.

, because I'm inferring that a backup system isn't being planned.

Because no matter which way you configure your primary storage system, RAID is not Backup. And if you really care about data, you will implement both.

 

[Arman]

If this statement is true:
then this statement confuses me: , because I'm inferring that a backup system isn't being planned.

Because no matter which way you configure your primary storage system, RAID is not Backup. And if you really care about data, you will implement both.

I can only spend so much money at a time... I will be saving money for an offsite backup or cloud solution in the future.

I never said RAID is a backup, nor I thought it was. However RaidZ3 is more redundant and therefor reduces the probability of a catastrophic data loss. Which does help my interests...

 

[Dice]

In case you are intrigued by the powers of magic redundancy of Raidz3, but do not have a solid backup solution (or a UPS if you are in regions which ever see power outages/fluctuating power net), you should <really> consider a raidz2 setup, save the money from the last drive, and as soon as possible get something like a shingled cheap huge (8? 10? TB) drive and put on a shelf (preferably off site) as a backup solution.
That would grant you <by far> better protection of your data than a Raidz3 vs Raidz2 setup would.

Cheers,
.

 

[depasseg]

you should <really> consider a raidz2 setup, save the money from the last drive, and as soon as possible get something like a shingled cheap huge (8? 10? TB) drive and put on a shelf (preferably off site) as a backup solution.
That would grant you <by far> better protection of your data than a Raidz3 vs Raidz2 setup would.

^^^ This

You are spending so much for additional drives so that you can lose 6 drives out of 10, which is just crazy. You are more likely at that point to be a victim of theft, fire, power surge, user error or whatever, that multi-RAIDZ3 vdevs will be irrelevant.

 

[Arman]

In case you are intrigued by the powers of magic redundancy of Raidz3, but do not have a solid backup solution (or a UPS if you are in regions which ever see power outages/fluctuating power net), you should <really> consider a raidz2 setup, save the money from the last drive, and as soon as possible get something like a shingled cheap huge (8? 10? TB) drive and put on a shelf (preferably off site) as a backup solution.
That would grant you <by far> better protection of your data than a Raidz3 vs Raidz2 setup would.

Cheers,
Dice.

Indeed, it's magical... :D
I did think of getting a UPS however in my location we get power-outages very rarely.
Yes you're right about getting a cheap external HDD for an offsite backup, however that does not encourage me to ignore the magical powers of RaidZ3. Just like RaidZ1 raid Z2 will become outdated sooner or later. I will buy an offsite backup asap after I put this whole thing together.

 

[Mirfster]

@Mirfster Couldn't I just rename the files? Wouldn't that cause them to be rearranged onto the new VDEV too? @Mirfster You did say that if the file is "modified" It would...

Nope, that would not really do the trick. Only way I can think of off the top of my head is to backup the data, delete it and then restore it.

If RaidZ3 is a "must" for you and you are looking at 10 drives, then simply do a single RaidZ3 vdev and call it a day. You will still get the redundancy of 3 drives for your pool but do not end up sacrificing 6 drives (if you went with 2 x 5 RaidZ3 vdevs).

Remember if you lose any vdev that is part of a pool; then the entire pool is lost. So having 2 x 5 RaidZ3 vdevs only gives the semblance of six drive redundancy, but that is from a "10 mile view". From a "ground level" perspective, your pool is only as resilient as the weakest link regardless of the strengths of the other fail-safes.

As others have said... backup, backup and backup... ;)

 

[Arman]

I love the FreeNAS community... Caring and helpful people :3

I'm gonna go spend some time thinking hard about my configuration. I'll update.

 

[depasseg]

I did think of getting a UPS however in my location we get power-outages very rarely.

Really? One of the simplest and most cost effective power protection devices should be another important piece of the equation for you to reconsider.

Nope, that would not really do the trick. Only way I can think of off the top of my head is to backup the data, delete it and then restore it.

Or just move it between 2 different datasets in the same pool. (you could even move it back if you wanted).

 

[Mirfster]

Or just move it between 2 different datasets in the same pool. (you could even move it back if you wanted).

Good suggestion.

 

[Arman]

Really? One of the simplest and most cost effective power protection devices should be another important piece of the equation for you to reconsider.

Or just move it between 2 different datasets in the same pool. (you could even move it back if you wanted).

Which UPS do you use? I'm considering buying one.

Also, in brief can you explain what datasets are and how I could move it between 2 datasets?

 

[depasseg]

I have a pair of APC's that feed my rack. It's important to size for the list that you are pulling and the duration you would like battery power to last.

Think of datasets as folders, you can move and copy between them. But they are also independent like partitions on a disk so you can configure them with different options.

 

[Arman]

@depasseg Would ~5 minutes be enough just shut off the system? How many VA(s) does your system pull?

I don't currently understand, I will come back to this once I get my NAS. However, is the thing you're saying difficult to do?

I have a question. The motherboard I will be using has 8 SATA ports. I want to connect 10 drives in total. Am I able to connect 8 drives via SATA and then get an HBA and connect the remaining two through PCI-e? Would that not cause any issues?

 

[Arman]

By the way, does FreeNAS 9.10 allow you to install it via USB on the new Skylakes? Or has that still not been resolved?

 

[Arman]

Out of curiosity: Say you set up a 5 disk VDEV configured in RaidZ2. Is it possible to change the configuration to RaidZ3 in the future if you wanted to (Provided you have not used over 40% of the available storage)?

 

[Dice]

Out of curiosity: Say you set up a 5 disk VDEV configured in RaidZ2. Is it possible to change the configuration to RaidZ3 in the future if you wanted to (Provided you have not used over 40% of the available storage)?

No.
You'll need to offload the data somewhere else and reconfigure the pool from scratch.

 

[Mirfster]

I have a question. The motherboard I will be using has 8 SATA ports. I want to connect 10 drives in total. Am I able to connect 8 drives via SATA and then get an HBA and connect the remaining two through PCI-e? Would that not cause any issues?

Nope, as long as FreeNas can get direct access to the drives then it really doesn't care which route it takes (one of the beauties of FreeNas). Only suggestion I would make is to know where each drive is located (physically attached). That way if you get a notification that two drives have failed from your RaidZ3 you can reference your document(s). It may turn out that those two "failed drives" are the ones attached to the HBA. In that case, it may just be the HBA itself instead of the two drives.

Example of an old "Hard Drive Matrix" of mine is in the attached image.

By the way, does FreeNAS 9.10 allow you to install it via USB on the new Skylakes? Or has that still not been resolved?

As far as I know that has been resolved with 9.10. However, I currently am not running any SkyLakes. Quick search in the forums should give you this info real quick though. *** I believe joeschmuck and Ericloewe are two contributors that are running SkyLake..

Out of curiosity: Say you set up a 5 disk VDEV configured in RaidZ2. Is it possible to change the configuration to RaidZ3 in the future if you wanted to (Provided you have not used over 40% of the available storage)?

Nope. Once the vdev is created the number of drives cannot be changed nor the RaidZ configuration. Only way would be to backup the data, delete, recreate and restore data. * Only speaking in RaidZ, not mirrors...

*** Edited to correct my statement.

 

[Arman]

I've been looking at power supplies for hours now and i'm now convinced that i'm looking at the correct ones... Are the connectors on server power supplies different to normal power supplies? If I get something like a EVGA SuperNOVA 750 P2 would I have any problems with it? Would it connect to the X11SSH-LN4F without issues? It also has 10 Sata connectors so it will be future-proof for me.
@danb35