× NETGEAR will be terminating ReadyCLOUD service by July 1st, 2023. For more details click here.
Orbi WiFi 7 RBE973
Reply

raid 6 factory default

TeknoJnky
Hero

raid 6 factory default

It has been awhile since I last factory defaulted, but tonight I was working on rebuilding and reorganizing my nas's, and this post is like a deja vu.

I have pro 6 running 6.8 beta 2 and factory defaulted with 6x 4tb drives.

I expected the default to raid-6 dual redundancy, however it kept defaulting to raid 5 (with 6 equal disks).

so my options appear to be that I had to reset again, using only 5 disks, switch to flex raid, wait for initial volume to sync, then I will have to add parity, wait for another resync and reshape before having a good raid-6 volume.

Alternatively I can destroy the default data volume, then recreate a new data volume with dual redunancy from the start, however doing this I lose all the default shares.

My understanding was that 6 disks+ was supposed to default to raid 6 ?

 

Model: ReadyNAS RNDP6000|ReadyNAS Pro 6 Chassis only
Message 1 of 62

Accepted Solutions
Skywalker
NETGEAR Expert

Re: raid 6 factory default

The documentation is incorrect.  I've issued a request to get it changed.  But a factory default will only give you a RAID 6 volume on systems with >6 drive bays.  It has always been this way.

View solution in original post

Message 5 of 62

All Replies
StephenB
Guru

Re: raid 6 factory default


@TeknoJnky wrote:

 

My understanding was that 6 disks+ was supposed to default to raid 6 ? 


That is what the manual says.  

Message 2 of 62
TeknoJnky
Hero

Re: raid 6 factory default

well as of 6.8b2 it is definately not defaulting to raid6 with 6 bay/6disks.

 

Message 3 of 62
StephenB
Guru

Re: raid 6 factory default

I forwarded this to development.

Message 4 of 62
Skywalker
NETGEAR Expert

Re: raid 6 factory default

The documentation is incorrect.  I've issued a request to get it changed.  But a factory default will only give you a RAID 6 volume on systems with >6 drive bays.  It has always been this way.

Message 5 of 62
btaroli
Prodigy

Re: raid 6 factory default

Can we submit an ER to request that change? Or provide the facility to request it during a factory reset? I ran into this during the last hardware test cycle, and it dumbfounded me. I got why it might not be the case with 4 bays, but with 6 or more and drives uptward of 4, 6, or 8 TB it gets a little crazy not using two parity stripes.

 

My understanding of XRAID was that it offered sensible protection without having to know much. But if I'm running 6x4TB or greater, it seems like the sensible thing is to run double parity. Could perhaps drive size be factored into the decision as well as number of drives?

Message 6 of 62
Sandshark
Sensei

Re: raid 6 factory default

It should definately not be the default for a 6-drive unit.  A delay period in which to make a choice, as there is in OS4.x, would be a good option.

Message 7 of 62
Skywalker
NETGEAR Expert

Re: raid 6 factory default

Primarily because opinions vary so widely, a lot of effort went into ReadyNAS OS 6 to make volume deletion and creation quick and simple, without requiring a factory default or even a reboot.

Message 8 of 62
btaroli
Prodigy

Re: raid 6 factory default


@Skywalker wrote:

Primarily because opinions vary so widely, a lot of effort went into ReadyNAS OS 6 to make volume deletion and creation quick and simple, without requiring a factory default or even a reboot.


 

I do recall, when doing a lot of testing of different drive combos (trying to figure out how to maximize space with a variety of disk sizes) the ease with which the data volume could be wiped without affecting the root volume. But I am left wondering -- since I wasn't paying a lot of attention to it at the time I did those tests -- what happens with the root volume if I factory default and re-create data as raid-6? Does the root volume (mdadm stripe) remain in raid-5? And if it does and I experience a double drive failure, where data volume would be safe, would the root volume fail? Or is root doing something even more sensible by default and maintaining itself as a striped mirror? 🙂

 

Ah, so it does...

 

# mdadm --misc --detail /dev/md0

/dev/md0:
        Version : 1.2
  Creation Time : Fri Dec 16 21:54:53 2016
     Raid Level : raid1
     Array Size : 4190208 (4.00 GiB 4.29 GB)
  Used Dev Size : 4190208 (4.00 GiB 4.29 GB)
   Raid Devices : 8
  Total Devices : 8
    Persistence : Superblock is persistent

 

Versus /data

 

# mdadm --misc --detail /dev/md127
/dev/md127:
        Version : 1.2
  Creation Time : Fri Dec 16 21:54:54 2016
     Raid Level : raid6
     Array Size : 23413000704 (22328.38 GiB 23974.91 GB)
  Used Dev Size : 3902166784 (3721.40 GiB 3995.82 GB)
   Raid Devices : 8
  Total Devices : 8
    Persistence : Superblock is persistent

# mdadm --misc --detail /dev/md126

/dev/md126:
        Version : 1.2
  Creation Time : Fri Dec 23 10:45:21 2016
     Raid Level : raid6
     Array Size : 11720636736 (11177.67 GiB 12001.93 GB)
  Used Dev Size : 3906878912 (3725.89 GiB 4000.64 GB)
   Raid Devices : 5
  Total Devices : 5
    Persistence : Superblock is persistent

Nevermind then! 😄

Message 9 of 62
Skywalker
NETGEAR Expert

Re: raid 6 factory default

The root volume is a RAID 1 mirrored across all the disks.  It won't be affected by volume deletion or degradation.

Message 10 of 62
jak0lantash
Mentor

Re: raid 6 factory default

https://community.netgear.com/t5/Using-your-ReadyNAS/XRAID-turned-RAID5-into-RAID6-when-adding-a-dri...

 

The simple rules for Volume Creation with X-RAID/OS6 (as per 6.5.0):
- Chassis with six bays or less:
1 HDD = JBOD
2 HDDs = RAID1
3-6 HDDs = RAID5

- Chassis with more than six bays:
1 HDD = JBOD
2 HDDs = RAID1
3-5 HDDs = RAID5
6+ HDDs = RAID6

Message 11 of 62
btaroli
Prodigy

Re: raid 6 factory default

 
Message 12 of 62
StephenB
Guru

Re: raid 6 factory default


@jak0lantash wrote:

https://community.netgear.com/t5/Using-your-ReadyNAS/XRAID-turned-RAID5-into-RAID6-when-adding-a-dri...

 

The simple rules for Volume Creation with X-RAID/OS6 (as per 6.5.0):
- Chassis with six bays or less:
1 HDD = JBOD
2 HDDs = RAID1
3-6 HDDs = RAID5

- Chassis with more than six bays:
1 HDD = JBOD
2 HDDs = RAID1
3-5 HDDs = RAID5
6+ HDDs = RAID6


Perhaps, but the manual for 6.7 just says 6+ HDDs = RAID6  It's a somewhat arbitrary threshold, personally I tihnk it should be the same rules for all chassis.

Message 13 of 62
btaroli
Prodigy

Re: raid 6 factory default


@StephenB wrote:

 

... I tihnk it should be the same rules for all chassis.

I tend to agree, partly for consistency and partly because with such large drives these days the rebuild times are getting long enough that I wouldn't ever want to run 6 without double parity. I can recall having drives fail on a 6-bay and watching it like a hawk during rebuild, worried about a secondary failure. I've actually seen this happen on ZFS with raidz1 and it sucks. Now with 8TB drives and larger ones yet to come, I'd never run them single parity. It just increases the risk. 

Message 14 of 62
StephenB
Guru

Re: raid 6 factory default

I think that when you are in XRAID mode the disk insertion behavior should always

  • let you choose redundancy or space (if extra redundancy is possible)
  • tell you what it is about to do (both initial setup and disk insertion) before it does it.  That means at least telling you the new volume size, and perhaps giving some guidance on extra space (e.g., "with dual redundancy you will need to upgrade 3 other disks before the volume can increase").
  • require your confirmation before it processes the new disk.

Somewhat less automatic to be sure, but it would prevent a lot of the expansion issues that show up here.

Message 15 of 62
jak0lantash
Mentor

Re: raid 6 factory default

I believe that anyone who wants a RAID6 should be well capable of building the NAS this way. X-RAID is made to remain simple.
That said, if you read my report I linked above, you'll see I agree that the rules and behavior are inconsistent.
For example, in a >6 bays chassis, you can have a 6 HDDs RAID5 volume with X-RAID (the reshape kicks in at the 7th HDD, not the 6th - as per 6.5.0)...

The simple rules for Horizontal Expansion with X-RAID/OS6 as per 6.5.0:
- Chassis with six bays or less:
1 HDD (JBOD) + 1 HDD -> RAID1
2 HDDs (RAID1) + 1 HDD -> RAID5
3-5 HDDs (RAID5) + 1 HDD -> RAID5

- Chassis with more than six bays:
1 HDD (JBOD) + 1 HDD -> RAID1
2 HDDs (RAID1) + 1 HDD -> RAID5
3-5 HDDs (RAID5) + 1 HDD -> RAID5
6+ HDDs (RAID6) + 1 HDD -> RAID6 (unless one-go reshape from RAID1 to RAID5)
Message 16 of 62
jak0lantash
Mentor

Re: raid 6 factory default

@StephenB
That's what I think Flex-RAID should do. Let you choose exactly what to do with unused capacity. The current options are very limited.
For example: add a secondary array for vertical expansion or extend the existing array. Multiple HDDs RAID1. Expand to RAID5 whenever, reshape to RAID6 whenever. Expand the existing array or add a new vdev for horizontal expansion.
In some situations, switching to X-RAID over and back may help achieve the goal, but that shouldn't be necessary.
Message 17 of 62
StephenB
Guru

Re: raid 6 factory default


@jak0lantash wrote:

That's what I think Flex-RAID should do. 

I agree XRAID needs to be simpler than Flex-RAID.

 

But over time, XRAID has become less automatic anyway.  OS 4.2 automatically formatted disks on insertion, OS 6 requires confirmation.  There are good reasons for that - too many users were inserting drives and losing data.  

 

I think from a human factors perspective, XRAID is still a bit too automatic.  Clearly too many users are caught by surprise by what it does, and once the volume is created their only recourse it to destroy it and restore the data.  IMO this can and should be fixed.

 

Documenting the rules isn't enough, since users generally don't see those rules until it's too late. 

Message 18 of 62
jak0lantash
Mentor

Re: raid 6 factory default

You have a point. And i do think that X-RAID shouldn't take the action immediately. Something like a 10 minutes timeout and/or manual confirmation with the explanation of what will happen and a simple question: continue, abort, switch to Flex-RAID for more options. would be a great improvement.
But I don't think X-RAID should give a choice of "options". That's Flex-RAID's job.
Message 19 of 62
TeknoJnky
Hero

Re: raid 6 factory default

thank you everyone for the additional info and thoughts.

 

My opinion matches some others, that I believe the rules should be consistent in that 6 disks should default to dual redundancy.

 

The ultimate goal of a nas is data security, and allowing those who know little about nas's to have a default mode that is less safe than it could be.

 

To quote Jack:

 

I believe that anyone who wants a RAID6 should be well capable of building the NAS this way.

 

I turn this around and say, anyone who wants a raid 5 with 6 or more disks, should be well capable of building the NAS this way.

 

To say it in other word, I believe that those who *ARE NOT* capable of deciding the most safe option, should have a default that is most safe for the data, and that is dual redundancy.

 

If a newbie user sticks 6 disks in his 6 bay nas, he should get dual redundancy.

 

If that user thinks, wow where is my space, then maybe the os6 GUI (like say during the setup wizard) should explicitly state that the default for 6 disks is dual redundancy to provide the most safe data storage, and provide a link to instructions on how to change this to other raid modes and point to a relevant KB article.

 

In summary, I think the rules should be consistent, and I think the default should be the safest option for those who don't (yet) know what to choose.

 

Message 20 of 62
btaroli
Prodigy

Re: raid 6 factory default

Some philosophical questions to be sure. I think that more recently (I've been using these for 10 years) there has been more of a small business/enterprise push. But XRAID's whole putpose was to ensure a USER with no knowledge of RAID could be functional by simply inserting disks. Most enterprise sorts would likely want more control and opt for FlexRAID.

When I consider XRAID, I think in terms of sensible default behavior that leads to a configuration that works and does a good job of protecting the unknowledgeable user. After all, what is user who arrives here typically asked to do? Download the logs and submit them for review. They are not expected to know how to fix things, or understand how they work.

So if XRAID were going to prompt for user input, are they going to have to go searching the web to understand what to do? That would not, IMO, be so good. Should they need to confirm that a disk inserted is ok to overwrite? Sure. Should they be asked to confirm or request a RAID level? Maybe not, unless it was done in a way that would make sense to them... such as prompting them (FP and/or email alert) to log in and then showing them, in the admin UI, the options in a way that makes sense to them. "Increase capacity OR increase redundancy?" And there should perhaps be a recommended option based on, for example, drive size/rebuild time.

In short, if XRAID evolves toward FlexRAID then important usability for unknowledgeable users is lost.
Message 21 of 62
TeknoJnky
Hero

Re: raid 6 factory default

I don't disagree with your comments, I just believe that default choices (in any application) is to provide the most sane and safe settings for the common or typical user.

 

And for me, that means dual redundancy for 6 or more disk (regardless of bay size).

 

this specific point does nothing to diminish xraid vs flex-raid.

 

I can understand your point about the slippery slope of making xraid more like flex-raid, but honestly the technology is the same, xraid now is simply auto-flex-raid, whereas flex-raid is manual-flex-raid, and if the most safe/usable defaults are used for both, then both should end up in the same place anyway. 🙂

 

 

Message 22 of 62
btaroli
Prodigy

Re: raid 6 factory default

I think I made clear earlier that I'm in definite agreement that I'd never do 6 without using raid6. But I can understand the perspective of a user who has 5 disks and has an expectation that adding a sixth will increase capacity. I'm not sure why that expectation would be different simply because their NAS has more slots though. 🙂

I also see prompting the user to make a decision based upon their action of inserting a disk as a compromise in automatic and default behavior. FlexRAID is and always should be completely manual.

I think one of the other posts mentioned thatbin some scenarios XRAID would ignore a disk insertion. What I'd hope is that the user would receive a notification such as "this disk cannot be automatically included in your volume due to its size..."
Message 23 of 62
TeknoJnky
Hero

Re: raid 6 factory default

Ok i see the confusion in our discussion;

 

I am specifically referring to factory defaulting with X number of disks.

 

It seems you are also including expansion behavior.

 

I think and expect the default expansion behavior to remain unchanged with x-raid, although I do believe that more information to the user regarding the last disk (and last chance to switch to raid 6) behavior, would be an improvement instead of blindly and irrevocably expanding the last disk without prompting the user.

 

IE something along the lines of, if new disk = final disk of %baysize% then prompt user: expand size or expand redundancy ?

 

I would be in favor of such a prompt, even though it takes a little bit away from 'xraid' automatic functionality, the option is just too important to leave to arbitrary default.

 

What I was previously trying to express is only the behavior of factory default/initial setup of 6 or more disks should be default to dual redundancy.

 

Message 24 of 62
jak0lantash
Mentor

Re: raid 6 factory default

Based on my experience, most complaints are about the lack of capacity, not the lack of redundancy. I agree that RAID6 is safer and somewhat more logical than RAID5 on 6 disks, but I think that most users want capacity, not dual redundancy, and most of them wouldn't know how or like to use Flex-RAID.
I saw a lot of disappointed users who couldn't understand why their volume capacity didn't increase even though they added a second drive...
Finally, many users would be better off having a single redundancy volume and an external backup, but that's another discussion.
Message 25 of 62
Discussion stats
  • 61 replies
  • 7911 views
  • 15 kudos
  • 10 in conversation
Announcements