AS-5110: Limited RAID 6 volume size or drive count?

[jchny]

Thinking of buying an AS-5110 with ten 4TB drives.

Trying to understand whether there is a limit of 8 drives for RAID 6 on the 5110...the migration table in Asustor College's "NAS 352 Online RAID Level Migration and
Capacity Expansion" Section 1.1 only goes up to 8 drives on RAID 6, so I'm wondering if that is a limit or not.

So, simple questions:

1. Can I build an AS-5110 with ten 4TB drives into a single RAID 6 volume? My alternative is to make two smaller RAID 6 volumes in the ten bays, with 4, 5, or 6 drives each, and lower total RAID 6 capacity because of the lower storage efficiency.
2. If I build a ten-drive RAID 6 array with 4TB drives and later somebody brings out 10TB drives, will I be able to run the migration tool to replace the 4TB drives with bigger ones?

Thanks in advance to anybody with a sure answer to those questions!

[crazynas]

Hi
1. Can I build an AS-5110 with ten 4TB drives into a single RAID 6 volume? My alternative is to make two smaller RAID 6 volumes in the ten bays, with 4, 5, or 6 drives each, and lower total RAID 6 capacity because of the lower storage efficiency. ->Yes, I think it's work.
2. If I build a ten-drive RAID 6 array with 4TB drives and later somebody brings out 10TB drives, will I be able to run the migration tool to replace the 4TB drives with bigger ones? -> Yes, I think it's work. due to AS5110T run x64bit kernel.

[Elrique64]

Thinking of buying an AS-5110 with ten 4TB drives.

Trying to understand whether there is a limit of 8 drives for RAID 6 on the 5110...the migration table in Asustor College's "NAS 352 Online RAID Level Migration and
Capacity Expansion" Section 1.1 only goes up to 8 drives on RAID 6, so I'm wondering if that is a limit or not.

So, simple questions:

1. Can I build an AS-5110 with ten 4TB drives into a single RAID 6 volume? My alternative is to make two smaller RAID 6 volumes in the ten bays, with 4, 5, or 6 drives each, and lower total RAID 6 capacity because of the lower storage efficiency.

Honestly I would set up a single or pair of disks as the system disks. These would be for where your apps are, any working HDD space you would need for them, etc. The remaining disks set up as a new/different volume and set up as Raid 6. Here's the documentation on Raids from Asustor. (It's not the most informative, but it's not too bad.)

http://download.asustor.com/college/en/ ... o_RAID.pdf

2. If I build a ten-drive RAID 6 array with 4TB drives and later somebody brings out 10TB drives, will I be able to run the migration tool to replace the 4TB drives with bigger ones?

Thanks in advance to anybody with a sure answer to those questions!

Asustor says that all of the NAS's can take 6TB drives. I can try finding the email from their support if you want/need it. Having said that, your question about upgrading the HDD's and Raid is going to depend on if the NAS itself can see a drive of the required size. It's speculation now, and I'm not going to even try guessing. I remember the 286 cpus limited to 500GB back in the day, but some MoBo's could be firmware flashed to allow higher cap drives to be on the system, you needed to run the larger drives with 2 or more partitions.

More information on migrating and increasing drive space can be found here:

http://download.asustor.com/college/en/ ... ansion.pdf

This is probably the best explanation of the process I've seen anywhere, so take this as a bible!

More pdf downloads for other various NAS functions can also be found here:

http://www.asustor.com/online/College?lan=en

The quality of the information here varies drastically. Some of it is really informative and others seem drastically lacking or just a brush over view. I suggested at one time that Asustor make all of these part of the NAS's integrated help system, but they have chosen (so far) to not do this. It is worth keeping a copy on the NAS itself I feel, so you can get a quick copy of something when you need it.

[jchny]

Crazynas and Elrique64,

Thank you for your answers.

I did not realize that I would need or should have a "system disk" in the AS-5110, but that makes sense. The OS may be in FLASH but I will need some space for apps and other files. Maybe I should get a couple of smaller hybrid SSDs like 2 or 3TB and configure them as RAID 1 array for the system volume, then configure the other eight bays as a RAID 6 array of 4TB drives for my main storage volume.

Multiple volumes within the eight bays might make sense but it will have an efficiency cost.

Once I've built the RAID 6 volume with 24TB of usable space, can I partition within that for access purposes, such as making one 16TB volume and two 4TB volumes?

[Elrique64]

Once I've built the RAID 6 volume with 24TB of usable space, can I partition within that for access purposes, such as making one 16TB volume and two 4TB volumes?

Uhmm... Short answer, no...

Long answer... the making of the raid makes the partitions, so that's all handled by the OS. If you want/need separate partitions you would need to do that by making the raid smaller by including less disks in the raid. Keep in mind, you might have more drive space than the OS can handle if you put in 8X6TB drives, even in Raid-6. The OS itself might not be able to see that large of a single volume. The OS might even have issues with 8X4TB. This is something you will have to try on your own. No one seems to have hard and fast numbers on what the OS itself can see on these newer NAS's. Even the info we DO have is only available for the older series of NAS's.

Yea! You get to play lab rat for this! Experiment to your heart's content and let us know the results... Word of warning though... Making a raid that large is going to take a while, so plan accordingly. If you try different drive configurations you could be at it for a week or more.

When you get the word on what the OS sees, rather than reply to this thread, make a new thread with your test results, if you would. It'll be less inclined to get lost in the ether that way.

[jchny]

Elrique64,

Thanks to you and the forum for your thoughts.

I've bought the 5110 and set it up with a two-drive RAID 1 array and a six-drive RAID 6 array.

So far so good. I'm a very happy customer. This is a great box and the software is quite friendly.

I'll post more in a new thread.

[Elrique64]

Glad I was able to help you. Remember, you have a lot more NAS than I do ATM, so I'm not going to be able to help with a lot, in the future. I'll do my best though..

[joelang1699]

Is having a system disk a requirement? Surely the system can be run from a 10 disk Raid 6 array other wise having a 10 bay NAS is effectively having a 8 or 9 bay system?

[orion]

Is having a system disk a requirement? Surely the system can be run from a 10 disk Raid 6 array other wise having a 10 bay NAS is effectively having a 8 or 9 bay system?

No, there is no requirement for system disk. After I ssh into NAS, the system locates actually on every disk (RAID-1). There are partitions for system. So, system is still running if we remove any of disks, but leaving only one disk. Of course, your data will be gone if you do the test .

[MikeG.6.5]

No, you don't NEED to have your system stuff and data stuff on separate disk volumes. The biggest reason I suggested it above, is from the sheer size of the disks the OP was planning to use. (Yes, that was my suggestion, even though different user name...)

We know there are limits to the OS's ability to access data, and with the 32-bit NAS's this limit is about 16TB. The 64-bit NAS's should be able to handle more, but Asustor hasn't been totally forthcoming on these figures. One would assume that it should be in the 32TB or higher range. But it's likely in the 24TB range.

I suggested to the OP that he set up a pair as a system disk array (Raid 1) to keep from hitting the maximum addressable HDD space limits of the OS, and to keep his data separate from the main OS functions. Things like torrenting and news grabbing can be hard on disks, if you have a lot going on. Putting the actual data on a separate volume is more of a fault tolerance and disaster recovery step than anything at that point.

Keep in mind, when you are using a raid array, if you lose one drive (raid 6 can handle 2 failed disks) when you put in a new drive, remaking the array is going to be harder on the remaining disks than most of the rest of the data manipulations you do. It's a potential weak point in the system. One drive fails, so you start rebuilding, another fails before that gets rebuilt. Now you have NO tolerance... If you keep going on the rebuild, another could fail and there goes your data.

Ideally, from a fault tolerance and data protection scheme, you would want a system volume of 1 or 2 disks (2 preferred) in Raid 1. 4 drives in raid (5 or 6) and the other 4 drives mirroring the first data array. This gives conceivably, 12-20TB of storage, mirrored locally, so if one drive fails, you can restore the whole thing either from the mirror or through rebuild. The final step for complete redundancy would be an identical NAS off-site with the same configuration, and having the data rsynced nightly to this second NAS, which handles the mirroring onto the second raid 5 or 6 array itself.

Keep in mind, this would be the ideal in an enterprise situation. Not a home user that just wants to store his media, or what have you. This would be the ideal if there is concern about privacy, and data integrity, and eliminates any cloud apps for the data storage.

Using the above method, with a system array and data arrays gives great fault tolerance should a single drive fail, allows you to rebuild the system without affecting your data, and allows you to be completely assured your data is maintained should anything happen to one or more disks in the system. This is NOT the norm, and usually only a recommendation with a 6-10 bay NAS. Those of us with smaller NASes obviously can't set this type of system up, so we have to go about doing it differently.

I have a 7004T (4 x 6TB in raid 5, roughly 16TB.) and a 202T (2 x 3TB in raid 1). The 202T has a 5 bay raid-5 USB enclosure on it (5 x 3TB, roughly 10TB total). Data is rsynced to it on a weekly basis from the 7004T, with that data going straight to the USB enclosure. Since the enclosure is smaller than the whole of the 7004T, I'm going to have to do some shuffling when things start filling up, but I have perfect off-site redundancy and fault tolerances built into the system. I can restore the whole of my data through a reversed rsync, should the need arise, and I know the data is good within the 7 day period. (I only lose the newest stuff, which can usually be rebuilt the easiest.)

HTH to clarify things a bit.