Extra redundancy


jonp

Recommended Posts

You can already mirror your parity drive by just using a RAID controller to create a mirrored volume; then assign that as the parity drive.

 

But a mirrored parity only provides extra protection for that ONE drive -- it doesn't provide any additional protection for your data, just for the parity drive.  So if you have a dual-drive failure, then you'll not lose any data IF one of those failed drives is one of the parity drives; but otherwise nothing changes.

 

Link to comment
  • 1 month later...
  • Replies 68
  • Created
  • Last Reply

Top Posters In This Topic

  • 2 months later...

The response to this request in the past has always been Limetech saying it would be a lot of work, and "helpful" folks on the forum arguing it's not needed. 

 

I don't know how much work it would be on Limetech's part to implement this, but to me at least it would seem like a better use of their time to develop, than many other features that have been added since this was first requested.  I'd even be willing to purchase an upgraded license version to get this feature.  A Pro-Advanced or a Super-Pro license maybe.  And I haven't given Limetech any money since 2008, but yet I've enjoyed the benefits of their work through several upgrade cycles for the past 7 years.  I'm wondering how many other people would be willing to pay for an upgrade like this as well.  A scenario like that might make it more worthwhile for them to spend their time on development, by adding a potential revenue stream from people who already purchased a pro license long ago. 

 

And the people who don't think it's needed, well they don't have to buy it. 

Link to comment

I'm wondering how many other people would be willing to pay for an upgrade like this as well.  A scenario like that might make it more worthwhile for them to spend their time on development, by adding a potential revenue stream from people who already purchased a pro license long ago.

 

I am, and I have purchased licenses for spare USB keys to help fund the current development stream even though I don't need them.

 

I suppose the proof is in the pudding.

People who want to fund the effort will need to make it known.

Link to comment

I'm wondering how many other people would be willing to pay for an upgrade like this as well.  A scenario like that might make it more worthwhile for them to spend their time on development, by adding a potential revenue stream from people who already purchased a pro license long ago.

 

I am, and I have purchased licenses for spare USB keys to help fund the current development stream even though I don't need them.

 

I suppose the proof is in the pudding.

People who want to fund the effort will need to make it known.

 

I have purchased licenses for spare usb keys that I have not redeemed in the many years I have been running unraid to fund development efforts. I also contribute significant time on the forums to help the community which contributes to their product.

 

There are many like Weebotech and myself who have already contributed significantly to fund the project in various ways, such as all the moderators, all the addon developers wether it be plugins or dockers, and all the addon service providers who support unraid.

Link to comment

The response to this request in the past has always been Limetech saying it would be a lot of work, and "helpful" folks on the forum arguing it's not needed. 

 

I don't know how much work it would be on Limetech's part to implement this, but to me at least it would seem like a better use of their time to develop, than many other features that have been added since this was first requested.  I'd even be willing to purchase an upgraded license version to get this feature.  A Pro-Advanced or a Super-Pro license maybe.  And I haven't given Limetech any money since 2008, but yet I've enjoyed the benefits of their work through several upgrade cycles for the past 7 years.  I'm wondering how many other people would be willing to pay for an upgrade like this as well.  A scenario like that might make it more worthwhile for them to spend their time on development, by adding a potential revenue stream from people who already purchased a pro license long ago. 

 

And the people who don't think it's needed, well they don't have to buy it.

 

I'd pay extra for something like that, for that extra little peace of mind.

Link to comment

I too would pay to have that additional feature.  Would I use it immediately?  Probably not unless I combined my two servers into one.  Currently the absolutely must have data is backed up on both servers, my main desktop, and DVD's.

 

But, with everyone now talking about paying extra, it might be worth it to note Tom's comment on this a while ago:  http://lime-technology.com/forum/index.php?topic=34427.msg323604#msg323604

 

That being said, the path which limetech has chosen (finishing off other nas core features) has to be finished first.  Otherwise we are going to wind up in a never ending beta cycle with nothing but half finished features. 

 

Perhaps a poll should be opened to try and gauge how many people would pay, and additionally how much extra they would pay

Link to comment

I'm wondering how many other people would be willing to pay for an upgrade like this as well.  A scenario like that might make it more worthwhile for them to spend their time on development, by adding a potential revenue stream from people who already purchased a pro license long ago.

 

I am, and I have purchased licenses for spare USB keys to help fund the current development stream even though I don't need them.

 

I suppose the proof is in the pudding.

People who want to fund the effort will need to make it known.

 

I have purchased licenses for spare usb keys that I have not redeemed in the many years I have been running unraid to fund development efforts. I also contribute significant time on the forums to help the community which contributes to their product.

 

There are many like Weebotech and myself who have already contributed significantly to fund the project in various ways, such as all the moderators, all the addon developers wether it be plugins or dockers, and all the addon service providers who support unraid.

I think you may have misunderstood the intent of my comments.  I want dual parity, and am willing to pay a reasonable amount of money to get it.  That's all I was trying to say. 

Link to comment

I think I would prefer to have multiple smaller arrays / pools instead of just dual parity. That way I could have 1 parity drive for every 4 data drives. This would show as a 4 array setup in a 20 drive system, with each segment (array/pool) capable of sustaining a single drive failure without losing of data.

Link to comment

I too would pay to have that additional feature.  Would I use it immediately?  Probably not unless I combined my two servers into one.  Currently the absolutely must have data is backed up on both servers, my main desktop, and DVD's.

 

But, with everyone now talking about paying extra, it might be worth it to note Tom's comment on this a while ago:  http://lime-technology.com/forum/index.php?topic=34427.msg323604#msg323604

 

That being said, the path which limetech has chosen (finishing off other nas core features) has to be finished first.  Otherwise we are going to wind up in a never ending beta cycle with nothing but half finished features. 

 

Perhaps a poll should be opened to try and gauge how many people would pay, and additionally how much extra they would pay

I agree completely, and I don't think they should drop the items that are in the current development pipeline to work on this.  But adding it to "next in line" in the pipeline wouldn't be too bad.  I'd pay up to $100 or so.  That's probably more than what other people would want to pay, but it's a feature I really want. 

Link to comment

I think I would prefer to have multiple smaller arrays / pools instead of just dual parity. That way I could have 1 parity drive for every 4 data drives. This would show as a 4 array setup in a 20 drive system, with each segment (array/pool) capable of sustaining a single drive failure without losing of data.

 

There is a disadvantage with this, 2 drives failing in the same array is likely to mean half your data is lost and at best 25% of your data is lost.  If you had dual parity across 20 drives, and 2 drives failed, you would not lose any data, additionally, you would be sacrificing only 2 drives rather than 4 for parity protection.  Statistically speaking, the likelihood of a drive failing is when it is most stressed, one of the times the drive is most stressed is when you are rebuilding a disk or doing a parity check.

Link to comment

I think I would prefer to have multiple smaller arrays / pools instead of just dual parity. That way I could have 1 parity drive for every 4 data drives. This would show as a 4 array setup in a 20 drive system, with each segment (array/pool) capable of sustaining a single drive failure without losing of data.

 

There is a disadvantage with this, 2 drives failing in the same array is likely to mean half your data is lost and at best 25% of your data is lost.  If you had dual parity across 20 drives, and 2 drives failed, you would not lose any data, additionally, you would be sacrificing only 2 drives rather than 4 for parity protection.  Statistically speaking, the likelihood of a drive failing is when it is most stressed, one of the times the drive is most stressed is when you are rebuilding a disk or doing a parity check.

 

No. Dont play with statistics. Percentages hide the real value.

 

A 2 drive failure within 1 pool still only means 2 drives of data is lost. A 4 drive failure where only 1 drive per pool fails means NO loss of data at all. That is an infinite times better than your situation of losing 4 drives in your dual parity setup which at best is losing of 2 drives of data with 2 drives of parity and at worse losing 4 drives worth of data.

 

Link to comment

I think I would prefer to have multiple smaller arrays / pools instead of just dual parity. That way I could have 1 parity drive for every 4 data drives. This would show as a 4 array setup in a 20 drive system, with each segment (array/pool) capable of sustaining a single drive failure without losing of data.

 

 

This is what I did. 

I remember having a 25 driver server and trying to lift it in a emergency.  No way will I do that again.

Now I have 4 servers, 3 micros 1 medium sized.

I also don't use drives that are too old.

Link to comment

I think I would prefer to have multiple smaller arrays / pools instead of just dual parity. That way I could have 1 parity drive for every 4 data drives. This would show as a 4 array setup in a 20 drive system, with each segment (array/pool) capable of sustaining a single drive failure without losing of data.

 

There is a disadvantage with this, 2 drives failing in the same array is likely to mean half your data is lost and at best 25% of your data is lost.  If you had dual parity across 20 drives, and 2 drives failed, you would not lose any data, additionally, you would be sacrificing only 2 drives rather than 4 for parity protection.  Statistically speaking, the likelihood of a drive failing is when it is most stressed, one of the times the drive is most stressed is when you are rebuilding a disk or doing a parity check.

 

No. Dont play with statistics. Percentages hide the real value.

 

A 2 drive failure within 1 pool still only means 2 drives of data is lost. A 4 drive failure where only 1 drive per pool fails means NO loss of data at all. That is an infinite times better than your situation of losing 4 drives in your dual parity setup which at best is losing of 2 drives of data with 2 drives of parity and at worse losing 4 drives worth of data.

 

I am sorry, but you are wrong.

 

The chance of a second drive failing during a rebuild is significant.  If you took the same 20 drive set up and had 4 disks as quadruple parity, it is *MUCH* less likely to lose any data than 4 pools of 4D+1P.  I am not proposing Quad Parity, but to take your example, 4 sets of 4D+1P is *MUCH* worse than 2 sets of 8D+2P.

Link to comment

The chance of a second drive failing during a rebuild is significant.  If you took the same 20 drive set up and had 4 disks as quadruple parity, it is *MUCH* less likely to lose any data than 4 pools of 4D+1P.  I am not proposing Quad Parity, but to take your example, 4 sets of 4D+1P is *MUCH* worse than 2 sets of 8D+2P.

 

I was going off the premise of having Dual Parity -or- MultiArray within the same server. Not having BOTH available. Of course if both was available, you'd have to be a fool not to select both.

Link to comment

I had suggested a crowdfund campaign, goFundme or something of that order.

My thought was, that might make it worth while and provide incentive.

I don't know what the logistics are of doing it that way, was just an idea.

 

Imagine that, guaranteed license sales for those who purchase via some campaign.

They're not keen on the idea at this point.

 

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.