RAID 0 vs 5 vs COR parity for home use

[ChalkChalkson]

In a recent post @SCHISCHKA and I had a disagreement about the importance of redundancy in home NAS applications. His point was mainly that you ought to have a proper backup anyway, so why bother with redundancy. Mine was that drive performance really only matters in niece applications, so RAID 0 is an unnecessary risk, and why not implement a redundancy system, one drive is a small price to avoid the potential issues and work of restoring from backup.

So I wanted to hear what you guys think about it, I really like to change my mind due to good arguments, so I am looking forward to hear some  

[ARikozuM]

Always have an offsite backup, an on-site backup or two (one on an eHDD and another on the server itself, if possible), and RAID 5 volumes so that any drive can fail while the array is rebuilt.

 

I prefer RAID 5 but I also use RAID 6 if the controller allows it.

[ChalkChalkson]

Just now, ARikozuM said:

Always have an offsite backup, an on-site backup or two (one on an eHDD and another on the server itself, if possible)

^^I think this is something everyone agrees with, well I think in most cases offsite backup + redundancy on site is probably enough for home use, but at least the backup part of it  

1 minute ago, ARikozuM said:

I prefer RAID 5 but I also use RAID 6 if the controller allows it.

I hate how unflexible RAID 5 is, when I get a random drive from a client I want to be able to throw it into my secondary NAS regardless of size 

[ARikozuM]

I don't know or care much about feuds, but in a home NAS redundancy and backups are both important.

You don't want to rebuild a disk every time a random bit becomes corrupt when running 24/7 operations and you don't want to rely on just a backup since that can take a good while to rebuild from. 

You also don't want to lose time that could have been used serving files to members of the household from losing one drive in the array.

 

As far as RAID 0, don't use it for NAS devices since you're going to get stopped by your 100Mb/s and 1Gb/s network speeds.

[ChalkChalkson]

2 minutes ago, ARikozuM said:

[...]

You don't want to rebuild a disk every time a random bit becomes corrupt when running 24/7 operations and you don't want to rely on just a backup since that can take a good while to rebuild from. 

[...]

As far as RAID 0, don't use it for NAS devices since you're going to get stopped by your 100Mb/s and 1Gb/s network speeds.

Yeah, that was pretty much my argument

[ARikozuM]

1 minute ago, ChalkChalkson said:

Yeah, that was pretty much my argument

I didn't see your post until now.

[ChalkChalkson]

1 minute ago, ARikozuM said:

I didn't see your post until now.

^^yeah, it is a bit buggy atm  

[Anghammarad]

In a NAS I would go Raid 1 in a 2 bay, and all 4 bay and more to raid 5. If there are a lot of bays (10 or more) and I need performance I would go Raid 10. 

 

But all raid isn't protecting from data loss. A backup should always be in place.

[brwainer]

My opinion is that redundancy is never necessary if you have good backups, but dear god I would not want to restore from my backups unless things have gone really catastrophically bad. Redundancy saves you from having to go through that process if only a single drive fails (or two if you've set that level of redundancy) while backups exist for restoring individual files from a certain point in time (e.g. old versions, or accidentally deleted, or you got infected by ransomware) and for complete restores when everything is going wrong.

 

As for types of redundancy, here's what I prefer:
RAID 1 / 10 - when using motherboard "hardware RAID", linux mdadm or similar, or Windows Disk Manager (not Storage Spaces); also when using ZFS or Storage Spaces if you don't have a ZIL/journal device, or when you are making a storage-optimized virtual disk

RAID5/RAIDZ1/Parity with 1 disk tolerance - when you have a hardware RAID card with onboard cache and a BBU, or ZFS and Storage Spaces if you have a sufficiently sized ZIL / Writeback Cache / Journal device(s)

RAID6/RAIDZ2/Parity with 2 disk tolerance - same as above, but when you are dealing with more than 8 disks in a single vdev or Virtual Disk

[Electronics Wizardy]

2 hours ago, ChalkChalkson said:

hate how unflexible RAID 5 is, when I get a random drive from a client I want to be able to throw it into my secondary NAS regardless of size 

well storage spaces parity and btrfs, snapraid, and other solutions like synology's there able to have parity raid with different disk sizes.

[ChalkChalkson]

Just now, Electronics Wizardy said:

well storage spaces parity and btrfs, snapraid, and other solutions like synology's there able to have parity raid with different disk sizes.

but these aren't true RAID 5, are they?

[Electronics Wizardy]

15 minutes ago, ChalkChalkson said:

but these aren't true RAID 5, are they?

No but operate on the same parity concept. Raid 5 and other normal raid levels make might more sense in the data center where the storage is often never upgraded and a new box is just bought. 

[ChalkChalkson]

3 minutes ago, Electronics Wizardy said:

No but operate on the same parity concept. Raid 5 and other normal raid levels make might more sense in the data center where the storage is often never upgraded and a new box is just bought. 

Well XOR as a checksum isn't really a true RAID 5 thing  

But I see where you are coming from, aside from the striping systems like unraid's are equivalent to RAID 5 (6)

And I have no issue with these (run one of them in fact)

[leadeater]

3 hours ago, ChalkChalkson said:

I hate how unflexible RAID 5 is, when I get a random drive from a client I want to be able to throw it into my secondary NAS regardless of size 

Well as long as the disk is equal or greater in size to the smallest disk in the array when it was created you can, disks don't have to be the same size in RAID it's just that they typically are.

[ChalkChalkson]

2 hours ago, leadeater said:

Well as long as the disk is equal or greater in size to the smallest disk in the array when it was created you can, disks don't have to be the same size in RAID it's just that they typically are.

but not to full capacity unfortunately  they get "cropped" to the same size 

[SCHISCHKA]

ok here is my argument. you do not need a home nas up 24/7. the time it takes to rebuild an array is the same as restoring from backup. redundant drive adds extra cost for what benefit? The cost is you are wearing out one drive that could be better used as in your rotating offline storage.

 

RAID protects you from:

drive failure

 

Using that redundant drive as a backup drive instead protects you from:

lighting

fire

theft

ransomware

user error

software bugs

AND drive failure

these are all real things that RAID cannot protect from.

 

About using RAID0. Iv been doing this for years and I have no problems with using RAID0. As long as you buy good quality drives it is no problem. Again not running 24/7 at home a good drive can out last your storage capacity requirements. RAID0 does not wear out a drive faster. How often does a drive fail? well once in its lifetime. So you are risking what exactly? one day for at sometime in the next ten years one of your drives will fail but it is ok because you have a proper backup right?

 

Don't waste your money implementing technology intended for business that loose money when their hardware goes down.

 

You do not have other redundant parts like PSU, you might not have hotswap bays, you prob do not have a UPS and there is unlikely to be any demand for your NAS when you are swapping out drives. Why waste money on a redundant part if you are also not going to get these other expenses too?

 

Its really simple if your data can fit on one drive then use one drive or use two smaller drives in RAID0.

Buy at least two other drives to use as backup. have your backup run every hour. Rotate your two backup drives every day and you will never use your data, and you never wasted your money on a drive you do not need.

[ARikozuM]

7 minutes ago, SCHISCHKA said:

ok here is my argument. you do not need a home nas up 24/7.

I would hate to have to turn my home server on whenever I'm going to use it. Assuming you don't mean parking and stopping drives fully after a set time. 

7 minutes ago, SCHISCHKA said:

the time it takes to rebuild an array is the same as restoring from backup.

Not when you have to wait for every drive to complete it's writes. 

7 minutes ago, SCHISCHKA said:

redundant drive adds extra cost for what benefit?

Being able to lose a drive in the case of a failure. Your server can still provide data while a disk is down/offline and while rebuilding a RAID array. 

7 minutes ago, SCHISCHKA said:

The cost is you are wearing out one drive that could be better used as in your rotating offline storage.

You should still have a backup for emergencies. Taking a drive out and rotating would be a great idea, but not everyone buys hot-swap bays or has rackmount (or cabinet) space reserved for labeling and storing HDD's (see Linus's bathroom server).

[SCHISCHKA]

9 minutes ago, ARikozuM said:

I would hate to have to turn my home server on whenever I'm going to use it. Assuming you don't mean parking and stopping drives fully after a set time. 

Not when you have to wait for every drive to complete it's writes. 

Being able to lose a drive in the case of a failure. Your server can still provide data while a disk is down/offline and while rebuilding a RAID array. 

You should still have a backup for emergencies. Taking a drive out and rotating would be a great idea, but not everyone buys hot-swap bays or has rackmount (or cabinet) space reserved for labeling and storing HDD's (see Linus's bathroom server).

you completely missing my point that a home NAS does not need to be available 24/7 and does not need systems meant for enterprise use. Does your home nas need to be running when you are sleeping or at work? You can set an alarm in BIOS to wake up and your OS should be able to let you set a shutdown time.

[brwainer]

47 minutes ago, SCHISCHKA said:

ok here is my argument. you do not need a home nas up 24/7. the time it takes to rebuild an array is the same as restoring from backup. redundant drive adds extra cost for what benefit? The cost is you are wearing out one drive that could be better used as in your rotating offline storage.

 

RAID protects you from:

drive failure

 

Using that redundant drive as a backup drive instead protects you from:

lighting

fire

theft

ransomware

user error

software bugs

AND drive failure

these are all real things that RAID cannot protect from.

 

About using RAID0. Iv been doing this for years and I have no problems with using RAID0. As long as you buy good quality drives it is no problem. Again not running 24/7 at home a good drive can out last your storage capacity requirements. RAID0 does not wear out a drive faster. How often does a drive fail? well once in its lifetime. So you are risking what exactly? one day for at sometime in the next ten years one of your drives will fail but it is ok because you have a proper backup right?

 

Don't waste your money implementing technology intended for business that loose money when their hardware goes down.

 

You do not have other redundant parts like PSU, you might not have hotswap bays, you prob do not have a UPS and there is unlikely to be any demand for your NAS when you are swapping out drives. Why waste money on a redundant part if you are also not going to get these other expenses too?

 

Its really simple if your data can fit on one drive then use one drive or use two smaller drives in RAID0.

Buy at least two other drives to use as backup. have your backup run every hour. Rotate your two backup drives every day and you will never use your data, and you never wasted your money on a drive you do not need.

Based on your point of view, there is nothing wrong with what you have said. However, there are certain assumptions you are making that would not be true for all people.

 

"the time it takes to rebuild an array is the same as restoring from backup" - If I have a drive die, I need to write 3-4TB to the replacement. If I had to restore from backups, I would have to write 12-16TB to the new array. For me, restoring from backups on my primary dataset would take 3-4x longer than repairing an array. And during a rebuild the array is still online and available, during a recovery from backups it is not. I may not be a business user, but I run services like Plex that other people in my household use often enough that I would annoy then during a recovery from backups. 24/7 uptime is not required, but a 24+ hour recovery would not be accceptable. 

 

"How often does a drive fail? well once in its lifetime. So you are risking what exactly? one day for at sometime in the next ten years one of your drives will fail but it is ok because you have a proper backup right?" - Each drive has a given likelihood that it will die "today". Based on your storage method, you can compute how likely it is that "today" your entire array will fail due to a failing drive, seperate from outside events like lightning that have different likelihoods. While RAID 0 doesn't increase the likelihood that individual drives will die, it does significantly increase the likelihood that on any given day you will lose the array, whereas redundant arrays significantly reduce that chance. While it is true that you can then just recover from backups, you come back to my first point, which is that it takes a long time to recover from a backup. Also, each drive may last years (Mine mostly last around 7 years) but in a RAID 0 array of 8 drives you are unlikely to go more than 150 days without a failure (7 year lifetime = 1/2555 chance a drive will die today = 0.039% chance for a single drive per day = 0.66% chance the array will fail today [8th root of 0.039] ~= 150 average array lifetime). The same math with a 10 year drive lifetime gives an average array lifetime of 157 days. And this is just the baseline average, based on luck and external factors you might have all your drives die much sooner or much later. Over an infinitely long time though, you would get that average. And I can't afford a 24+ hour rebuild twice a year.

 

It sounds like you are talking about a smaller home NAS, since you talk about all your data fitting on one or two drives. In that situation I can't fault your advice overall. I have a techie friend who barely has 3TB of total data on their NAS, and has a 1TB RAID1 and a 2TB RAID0. For offsite backups they use a single drive for both volumes. The reason why there is such a discrepency between what you are talking about and what myself and others who are responding in this thread are talking about is that we have vastly more data, and many of us also want to use some business-type techniques because we are enthusiastic about them and want to practice them at home, on what is effectively a "home lab".

[brwainer]

23 minutes ago, SCHISCHKA said:

you completely missing my point that a home NAS does not need to be available 24/7 and does not need systems meant for enterprise use. Does your home nas need to be running when you are sleeping or at work? You can set an alarm in BIOS to wake up and your OS should be able to let you set a shutdown time.

In my household, someone is up almost 24/7, so shutting down the NAS wouldn't really be an option. I get your point, for the average perspn that is true, but this is a forum for tech enthusiasts for whom the arguments you are making do not apply.

 

EDIT: As for work, I work from home, so I'd like it to be online 24/7. And many people on this forum run services like Plex or OwnCloud that they may want to access remotely on their lunch break.

[SCHISCHKA]

7 minutes ago, brwainer said:

It sounds like you are talking about a smaller home NAS, since you talk about all your data fitting on one or two drives. In that situation I can't fault your advice overall. I have a techie friend who barely has 3TB of total data on their NAS, and has a 1TB RAID1 and a 2TB RAID0. For offsite backups they use a single drive for both volumes. The reason why there is such a discrepency between what you are talking about and what myself and others who are responding in this thread are talking about is that we have vastly more data, and many of us also want to use some business-type techniques because we are enthusiastic about them and want to practice them at home, on what is effectively a "home lab".

Yes i am talking about a HOME nas not a business nas. I dont know how many times i have to parrot myself on that

[brwainer]

Just now, SCHISCHKA said:

Yes i am talking about a HOME nas not a business nas. I dont know how many times i have to parrot myself on that

For many of the people on this forum, and for the ones who have thus far commented on this thread, our home NASs have way more than 2 drives worth of data. You cannot categorically say that my system is not a home NAS. What you are saying makes sense and is true for small home NASs, but the way you are arguing against redundancy does not universally make sense for all "home NAS" units.

[SCHISCHKA]

1 minute ago, brwainer said:

In my household, someone is up almost 24/7, so shutting down the NAS wouldn't really be an option. I get your point, for the average perspn that is true, but this is a forum for tech enthusiasts for whom the arguments you are making do not apply.

In my household we have jobs and we sleep. I would not make high assumptions about the people on this forum. I would say most are students. When I want to discuss technology at a level above retail I go to the spiceworks forum.

[brwainer]

2 minutes ago, SCHISCHKA said:

In my household we have jobs and we sleep. I would not make high assumptions about the people on this forum. I would say most are students. When I want to discuss technology at a level above retail I go to the spiceworks forum.

I will grant that I don't know about most of the people on this forum, I just stay in the "Servers and NAS" and "Networking" sections. So I have to amend what I said about people on this forum to mean the people who are active participants of these two sections of the forums. And I will point to the Storage Rankings thread on this forum that has 120 systems with an average of 31.4TB Raw disk space per server. Most of the active participants of these sections have systems on this list. So when you make a statement about small NASs, it really doesn't apply to the active people who are most likely to read and respond. I am not saying this to disuade you from your point, and I do not think you should stop bringing this point up whenever someone talks about a small NAS. I just think it explains why you get the response that you do when the point is raised, and this resulting thread.

 

I don't commonly put much effort into people's posts that are asking about small NAS systems, and I certainly don't see the ones that are posted outside of this section of the forum. For ones that I see in the future, I will still advocate for some level of redundancy just because it is more convenient to repair an array than to have to recover it. However, if the tone of the responses indicate that everyone is completely against having a single drive or RAID0, then I will certainly do what I can to make the topic as helpful to the OP as possible. There are certainly people for whom recommending RAID1/5/6/etc is overkill and for whom that is bad advice. My goal is to always be as helpful as possible when giving advice, because that is where I gain my pleasure from this forum.

[scottyseng]

9 hours ago, ChalkChalkson said:

-snip-

Maybe it's just me, but I prefer having redundancy. I got RAID6 on my two arrays, a primary (10 x 4TB), and a secondary (6 x 4TB). It's saved me several times already. I had a bad batch of drives where on my primary array I had two drives die while and to RMA both and hope another drive didn't die (Though paranoia got the better of me and I went and bought a new drive to rebuild the array while waiting on RMA). The smaller secondary array had two drives get rejected because I messed up and had one of the two SAS cables tilted in the connector and not in all the way...so the RAID card thought the drives were bad. Both array are rebuilt and are just fine today.

 

I say get redundancy because it takes a really long time to have to initialize a new array vs rebuilding. The other major thing for me is that I can use my array while it's rebuilding, not so much when I have to recover from a backup.

 

I remember for me to build the initial primary array (6 x 4TB at the start), was around 12 days 24/7. To rebuild one drive took around 4-5 days. Also have to consider the time it would take to move the data from the backup to the new array as well, and I have a lot of data at 20TB (Though I'd say about 1/3 is critical data...rest I can rerip from the physical media I own or download from where I bought digital media from).

 

The other plus of having RAID5 / 6 on traditional RAID is that you can expand the array, sure it takes a long time (The jump from 9 x 4TB to 10 x 4TB took 20 days...), but it's better than having to buy drives and be stuck with them after the new array is built just to temp store your stuff. I used originally have a RAID10 (4 x 4TB), but I had to actually buy two drives and put them in RAID0 as a temp location to move my data to as the new array was built. Ever since changing to RAID6, I've never had any downtime at all, especially when needing to expand my NAS arrays.

 

Also, my NAS is pretty much on 24/7 because even running a defrag takes several days...I'm also not home (work / college), but the NAS pretty much just idles or I give it something to render. I'm also the only one that uses the NAS. I don't even use it to stream anything to any TVs or media PCs. It's literally just a really large drive for my main PC (It sadly has better hardware than my main PC as well).

 

It definitely depends on the user case though. Not everyone needs a huge home NAS. However, if the NAS is going to use three drives anyway or four, I'd push for RAID5 or 6 just for that extra bit of protection.