SOLVED raidz1 + stripe?

[Markes12344]

So i have 5x4TB WD red drives now.
i just finished replacing my 5x1TB Black n Blue Drives.

When I created the Raidz1 I did so with 4 drives in Raidz1 and 1 drive as a spare, and my small 60GB SSD as Cache.
but then I realized during the HDD swap I could lose 2 Drives and still have my files, therefore I realized I didn't need my Spare.

So I deleted my spare and the only way to re add that HDD is to do a manual setup in volume manager. Now that drive is as a Stripe not part of the Raidz1, but my file size did go from 11.5TB available to 15TB available (yes i already have 3TB in the array due to it being part of my existing NAS)

So my question is this,

Whats the difference in the drive being labeled as part of my Raidz1 vs Stripe?

 

[nojohnny101]

You cannot add a drive to a vdev (except mirrors).

So what you have right now is a single drive that is "striped" to your existing raidz1 vdev. If that stripe fails, you will lose your entire pool.

 

[Markes12344]

say what!k so im gunna back up delete n remake this raidz..
what is a stripe in this regards?

 

[nojohnny101]

If you have a place you can temporarily store you data, yes you should backup your data and then remake your vdev. If you don't have a special usage case and just want a blend of performance and redundancy (which is not a substitute for a good backup plan) then raidz2 might be good choice.

You should check out @cyberjock and his presentation on the basics of freenas including detalied and concise explanations on zpool, vdev, striping, ecc ram, etc.
https://forums.freenas.org/index.ph...ning-vdev-zpool-zil-and-l2arc-for-noobs.7775/

 

[Markes12344]

so is there a simple explanation of what the stripe drive is? just for future reference? and i'm just going to be creating a raidz1 again, but with all 5 drives.

 

[nojohnny101]

No offense but it seems you didn't even read what I posted. I already told you what I recommend you rebuild as (raidz2). I also posted a link that answers your second question in much greater detail then I could.

Most of the questions you are asking are pretty basic, and most people on here expect you to do research first then ask questions. If you go through the slideshow and don't understand, then post back. We are here to help, but not here to do your own reading for you.

 

[Markes12344]

i briefed though that slide, sorry not here to make you guys do the reading but i figured someone would just know... from what i found members of the raidz1 have the information split amongst them and some info will be shared amongst them so when one or two dies all information is saved, and adding the stripe is simply put as another drive for the information to be split amongst but won't have some information shared between the remaining drives in the Raidz... So I thinks that's my answer.

You suggest Raidz2 vs Raidz1, is there a short answer to that?

 

[gpsguy]

It provides more redundancy for your pool.

Here's some more information: http://lmgtfy.com/?q=RAIDz2+vs+RAIDz1

You suggest Raidz2 vs Raidz1, is there a short answer to that?

 

[Stux]

A bit like having a spare which does something.

I use 5x3TB in RaidZ2 in my backup system.

 

[Markes12344]

A bit like having a spare which does something.

I use 5x3TB in RaidZ2 in my backup system.

Thanx for the simple...

like I stated I essentially removed two drives from raidz1 (so 2/4raidz1 + 1/1spare +1/1cache) and yet I was still able to recover all data. did i get away with it because of my spare? The way I read the stuff on other threads raidz1 only allows one drive fail while raidz2 allows two drive fail.

 

[Stux]

Raidz1 only allows one drive failure. And if you have one more minor error you could lose the array.

The cache and spare drives were not contributing to redundancy. Having a "spare" is a waste, it should've been a 5 disk raidz2 instead.

Raidz2 allows two drive failures, but more importantly after the first drive failure, you are still protected against minor errors.

 

[danb35]

Raidz1 only allows one drive failure. And if you have one more minor error you could lose the array.

Possible, but highly unlikely. Metadata is stored in multiple copies in different places on the disk, and the more critical the metadata, the more copies are kept. To result in complete loss of the pool, every copy of the uberblock would need to be corrupted, and IIRC, there are six copies. A read error in the data will result in loss or corruption of that data, though--but at least ZFS will know about it and tell you.

 

[Markes12344]

Raidz1 only allows one drive failure. And if you have one more minor error you could lose the array.

The cache and spare drives were not contributing to redundancy. Having a "spare" is a waste, it should've been a 5 disk raidz2 instead.

Raidz2 allows two drive failures, but more importantly after the first drive failure, you are still protected against minor errors.

Thanx again, ya so it's weird that I was able to remove two, I was doing the drive conversion one at a time as I thought raidz1 was only a 1drive fail safe, yet when I disconnected the seconds power yet still recovered the volume even at the 2/4 remaining (well with one 4TB drive in place of its 1TB drive so 3/4 drives in the raidz1)

 

[pirateghost]

Thanx again, ya so it's weird that I was able to remove two, I was doing the drive conversion one at a time as I thought raidz1 was only a 1drive fail safe, yet when I disconnected the seconds power yet still recovered the volume even at the 2/4 remaining (well with one 4TB drive in place of its 1TB drive so 3/4 drives in the raidz1)

Removing a spare and a cache disk didn't do anything to the raidz because they aren't part of the raidz.

 

[Markes12344]

Removing a spare and a cache disk didn't do anything to the raidz because they aren't part of the raidz.

it was not the spare or cache drive i removed, not sure how the array managed to live, if it was to fail i would just re add a disk i removed obviously, but either way it is not the main point to my question, i guess i should of marked as solved. thank anyways.

3TB transferred to my main rig in roughly 8-9hours total.
Wiping the 5x4TB drives and creating a Raidz2 amongst them

 

[nojohnny101]

Wiping the 5x4TB drives and creating a Raidz2 amongst them

Great!

it was not the spare or cache drive i removed, not sure how the array managed to live, if it was to fail i would just re add a disk i removed obviously, but either way it is not the main point to my question, i guess i should of marked as solved. thank anyways.

well that isn't technically possible, unless you're a wizard. are you a wizard? ;)

raidz1 only survives 1 drive failure period. you're on the right path now and will be much better protected going forward especially given the size of drives you are using (>2TB).

 

[Markes12344]

Great!

well that isn't technically possible, unless you're a wizard. are you a wizard? ;)

raidz1 only survives 1 drive failure period. you're on the right path now and will be much better protected going forward especially given the size of drives you are using (>2TB).

My sourcery is a mystery to all., how are you not part of all?

The sad part is, I need to budget in another drive.. Needed 15TB for my house holds one stop shop for their files, although its a gamble whose set up will fail first., this server or their PC
But I think I may as well wait a year so I can just go ahead n make a new system that can unraid the system n have a freenas/WinOS system so I can game server host. It's too bad I had to sacrifice my old one this time around for the Freenas.