A few cases out there, like the Norco 4224 and Supermicro CSE-846TQ-R900 have 24 bays. As a 4224 owner with only 2 logical disk slots left, yet 5 physical slots available, would it be possible to include 3 more logical data slots to max out our arrays in the next beta(s)? Having 1 parity and 23 data drives to fill these servers would be a very welcome feature.

A few cases out there, like the Norco 4224 and Supermicro CSE-846TQ-R900 have 24 bays. As a 4224 owner with only 2 logical disk slots left, yet 5 physical slots available, would it be possible to include 3 more logical data slots to max out our arrays in the next beta(s)? Having 1 parity and 23 data drives to fill these servers would be a very welcome feature.

 

Yes this will get on the Roadmap.  It's far better though to have 24 array drives: P-drive, Q-drive, 22 data-drives.  Then any 2 drives can fail and it's possible to recover data.

 

A few cases out there, like the Norco 4224 and Supermicro CSE-846TQ-R900 have 24 bays. As a 4224 owner with only 2 logical disk slots left, yet 5 physical slots available, would it be possible to include 3 more logical data slots to max out our arrays in the next beta(s)? Having 1 parity and 23 data drives to fill these servers would be a very welcome feature.

 

Yes this will get on the Roadmap.  It's far better though to have 24 array drives: P-drive, Q-drive, 22 data-drives.  Then any 2 drives can fail and it's possible to recover data.

 

 

I do agree with that, and it sounds like a great plan. Thanks for the update and best of luck with the beta's!

 

Also asking for some priority on this.  My 4224 is feeling sorry for its missing drives

I could do with 1 more data drive at least -- I have space for 22 drives in my case and no need for a cache drive.

Also asking for some priority on this.  My 4224 is feeling sorry for its missing drives

I could do with 1 more data drive at least -- I have space for 22 drives in my case and no need for a cache drive.

 

I wouldn't mind 2 more, as my 4224 is almost maxed out    That would still allow one slot to be open for dual parity later down the road.

The only saving grace for unRAID is the fact that losing two drives simultaneously won't nuke all your data. Because of this, I can let slide the fact that it only supports 1 parity drive (ala RAID5).

 

I would never run a 20 drive RAID with only 1 parity in my professional career. I hope dual parity does make it to the top of the list soon as I plan on growing my array quite substantially in the near future.

The only saving grace for unRAID is the fact that losing two drives simultaneously won't nuke all your data. Because of this, I can let slide the fact that it only supports 1 parity drive (ala RAID5).

 

I would never run a 20 drive RAID with only 1 parity in my professional career. I hope dual parity does make it to the top of the list soon as I plan on growing my array quite substantially in the near future.

 

Just as well if this was a professional environment you wouldn't have just one server you would have a couple at least.

The only saving grace for unRAID is the fact that losing two drives simultaneously won't nuke all your data. Because of this, I can let slide the fact that it only supports 1 parity drive (ala RAID5).

 

I would never run a 20 drive RAID with only 1 parity in my professional career. I hope dual parity does make it to the top of the list soon as I plan on growing my array quite substantially in the near future.

 

Just as well if this was a professional environment you wouldn't have just one server you would have a couple at least.

 

And the whole thing would be backed up regularly too!

You'd be surprised how often that doesn't happen.

 

Budgets.. cutbacks...

 

But yeah, I think dual parity is important to add.

Does dual parity protect us from 2 disk failures in unRAID in the way it does in RAID6? Sure if we lose both parity disks we are good, or on parity disk and one data disk, but if we lose 2 data disks, we have 2 copies of our parity and a hole in our array. Or am I missing something?

When implemented, dual parity (Tom usually refers as P+Q parity) will protect you from 2 simultaneous disk failures - including 2 failed data disks. It will not just keep 2 copies of the parity disk.

 

Note this feature does not exist at this time. 

When implemented, dual parity (Tom usually refers as P+Q parity) will protect you from 2 simultaneous disk failures - including 2 failed data disks. It will not just keep 2 copies of the parity disk.

 

Note this feature does not exist at this time.  

 

 

that's just BRILLIANT!!!... can wait for this feature...

great to hear there are plans for an extra parity drive. it has helped me decide to jump on the Unraid train. now all i need to do is think out a build ^^

Bumping, in hopes that we can soon get 3 more slots for data drives. I just filled the last two slots in my array, maxing it out, yet I have 3 more physical slots unused in my 4224..

is the current limit 20 data drives? i have 20 data drives in my Norco 4020 plus a cache and a parity (these are both mounted above the hotswap bays) giving me a total of 22 drives. I think that this is enough since even at current limits i could have 40TB of data available (i only have 21TB right now) and with the addition of 3TB support that will take me up to 60TB potentially. i know that people have lots of drives sitting around that they would like to use, but what are the costs of running another 20 drives? just because you have them doesn't mean that you have to use them. i would rather run (1) 2TB drive than (4) 500GB drives.

 

just my opinion, but if there is a way to add more than 20 data drives i would probably do it too just because i have the extras laying around. 

 

kingpin

I'm at 11 drives right now and would love to see a second parity drive.  I love technology, it just doesn't love me back.  I could see myself being the first person to put the dual recovery to use. 

I am pretty new here. I am looking for a new OS for my Norco 4242 (WHSv1 must GO!). Unraid is looking like a viable candidate.

 

A double parity drive would sell me on it. I know people are all hyped about the 3tb HDD support, but in the real world, drives fail. Also the more drives in the array, the more likely of a multi-drive disaster.

 

I work in an enterprise environment and more then once i have had an array (HP and Dell) spit out a bad drive. in rebuilding the array (manually reinserting a drive or from a hot spare) the array be lost because a second drive takes a dump in the rebuild in a raid 5. whether the drive was having bad reads all along due the raid5 write hole or just the rebuild stress put it over the top to go tits up. As we are migrating to larger arrays (both drive size and density) no longer use raid 5 on any production servers. they are all hardware Raid 6, 10, 60 or equivelent ZFS pools depending on the need.

 

also when building a new array with non enterprise drives, we have more then once had several go bad in the first 90 days. sometime within days/hours of each other.

 

While my initial use for my unraid server is not going to be the only storage point for my files. it would be nice know i can have 20+ 2TB (or larger) drives in my array and have some peace of mind knowing that i can rely on unraid to keep data intact in a HDD disaster, but also knowing i wont have to spend a week restoring the drives.

A double parity drive would sell me on it. I know people are all hyped about the 3tb HDD support, but in the real world, drives fail. Also the more drives in the array, the more likely of a multi-drive disaster.

This also makes an argument in favor of 3TB drive support, because your quantity of drives is 33% lower for the same capacity as 2TB drives.

 

I work in an enterprise environment and more then once i have had an array (HP and Dell) spit out a bad drive. in rebuilding the array (manually reinserting a drive or from a hot spare) the array be lost because a second drive takes a dump in the rebuild in a raid 5. whether the drive was having bad reads all along due the raid5 write hole or just the rebuild stress put it over the top to go tits up.

One of the unique advantages of unRAID's RAID 4-ish model is that losing 2 drives doesn't kill the array, it only loses those two drives.  So even without P+Q parity, unRAID arrays fail much more gracefully than other "enterprise" RAID configurations.

 

also when building a new array with non enterprise drives, we have more then once had several go bad in the first 90 days. sometime within days/hours of each other.

This is a primary reason why many of us run our drives through 2 or 3 preclear cycles prior to trusting them.  This gives us the best chance of finding problems with the drives well in advance of an actual failure.

 

it would be nice know i can have 20+ 2TB (or larger) drives in my array and have some peace of mind knowing that i can rely on unraid to keep data intact in a HDD disaster, but also knowing i wont have to spend a week restoring the drives.

This is a bit confusing.  I've never heard of recovering from a drive failure without having to do a restore of some kind.  The larger the disk, the longer the restore will take.  However, I can say my rebuilds in simulated drive failures for my 2TB drives take around 9 hours.

 

This also makes an argument in favor of 3TB drive support, because your quantity of drives is 33% lower for the same capacity as 2TB drives.

 

 

Eh?   Don't get me wrong.  I understand the mathematical logic.   But in the 'real' world, or at least mine,  I have already invested in drives.   And I'm not about to toss away all my 500, 750's or 1000's...    I would love to see P+Q first...  but I know folks want the 3t support, so I don't make a big deal about it right now.

 

I don't think the 33% logic works for existing servers.

 

Personally, I'd do P+Q in an INSTANT to protect my 9 current drives, and soon to be (once I move a little more data from WHS over) 14 drives.

 

This also makes an argument in favor of 3TB drive support, because your quantity of drives is 33% lower for the same capacity as 2TB drives.

 

 

Eh?   Don't get me wrong.  I understand the mathematical logic.   But in the 'real' world, or at least mine,  I have already invested in drives.  

 

Not everyone has the horsepower to handle P+Q either, so there is a compromise required by some people no matter which feature is developed first.

 

I don't think the 33% logic works for existing servers.

 

You're correct, it doesn't make near as much sense if you're talking about existing infrastructure.  However, I assumed this is not the case with Johnm's new build, based on this...

...it would be nice know i can have 20+ 2TB (or larger) drives in my array...

 

Not everyone has the horsepower to handle P+Q either, so there is a compromise required by some people no matter which feature is developed first.

 

 

No offence meant, but the 'horsepower' to do P calcs, and P+Q calcs is trivial. 

 

http://lime-technology.com/forum/index.php?topic=12618.msg120052#msg120052

 

CPU speed does not affect parity calc speed as any modern CPU is more than fast enough to handle the minuscule load added by a parity calc.  Many pocket calculators could probably handle it 

 

No offence meant, but the 'horsepower' to do P calcs, and P+Q calcs is trivial.  

No offense taken, but Raj is assuming a Reed-Solomon calculation pattern, and there have been discussions over the past year about diagonal parity instead.  Lime-Tech initially indicated an inclination towards Reed-Solomon, but I think the P+Q Project should really just be called "Dual Parity" for the time being.  Raj's comment is very accurate, but it assumes one method for dual parity calculation, and there are in fact several for Lime-Tech to choose from.

 

There have also been discussions about data integrity checks (bit rot correction, resilvering, etc), and finding some way to integrate those with a dual parity scheme.  That was my what I was actually referring to, although I obviously didn't make that clear.

No offence meant, but the 'horsepower' to do P calcs, and P+Q calcs is trivial.  

No offense taken, but Raj is assuming a Reed-Solomon calculation pattern, and there have been discussions over the past year about diagonal parity instead.  Lime-Tech initially indicated an inclination towards Reed-Solomon, but I think the P+Q Project should really just be called "Dual Parity" for the time being.  Raj's comment is very accurate, but it assumes one method for dual parity calculation, and there are in fact several for Lime-Tech to choose from.

 

There have also been discussions about data integrity checks (bit rot correction, resilvering, etc), and finding some way to integrate those with a dual parity scheme.  That was my what I was actually referring to, although I obviously didn't make that clear.

 

I'll think that I've let this thread go too far on the side-track.      Thanks for the details.   The math makes my brain hurt.

I'll think that I've let this thread go too far on the side-track.      Thanks for the details.   The math makes my brain hurt.

 

I agree, we went pretty OT there.  Just so you know, the math makes everyone's brain hurt, even the seasoned guys.