You know what, I have been researching to having a solution for iCloud and Google One Drive because I am really struggling saving my image RAW files and videos. Then all on a sudden a photographer Tony Northrup brought the light of a NAS! I did not know what NAS is until couple of weeks ago! Then I started to do my own research and found you. I know you dont have smooth voice and attraction catching vocal gestures, but I find myself in you, I would want to express my research so that people can decide what's best for them. I have found the same agony in you. You are like a tech big brother who wants to advice whats best for us instead biased brand marketing. I like your videos. Just wanted to pay my gratitude because I know, a small wish can boost up the moral energy a lot cause you have done so much research, night and day sleepless time. I know for the video but I know it's for the people whom you want to help so desperately. Thank you so so much.
The time to rebuild an array with 1 of several smaller failed drives verses the time it takes to rebuild an array with 1 of two large drives is important to me as a home user.
I think it really depends where you are and what you can get. In some countries, it's cheaper to get bigger drives and a two bay NAS, where other places, smaller hard drives and a four bay NAS ended up being the better option. I've always tackled it as a your-mileage-may-vary scenario. Although if you want maximum capacity, a bigger NAS is the way to go.
"I think it really depends" could easily be the sub title of this video tbh. All of the points I touch on here are general and "more often than not" categories. I guess the main thrust of it is that there is no definitive answer to Big drive v Many drives, but there are enough smaller points to make one argument more compelling than the other for different users. Personally...I carry 87 USBs in a carrier bag and I shan't have it any other way....
Add price to this whole inflation, cost of living everything going up, it basically limits people .. "You'd like to get a large drive, if you could only afford it"
@@Tech-geeky Well, prices being what they are now, you'll really need to have some abnormally large amount of data for some reason, to get to where you're economically limited realistically. Second hand HDD's cost very little, have enough of them on hot-spare and cold-spare and you can run your array on a potato PC from 10yrs ago, that you have a pair of mirroring each other. Boom, Bob's your uncle - nearly nukeproof. Heck, if you want to be nukeproof, get a cheap laser printer and print the data out into a fire proof bin. Most things can be done on very limited budgets with time and effort, which you ought to have if your budget is severely limited. In most cases people have at least a lot of either time or money, both can solve these sorts of things, just different ways.
Really interesting video, thanks. Here are a couple of ideas on presentation of the numbers: 1. Produce a line graph of price against drive size. That would make it clear that the price per terabyte drops a lot from 1TB to 20TB drives - $35 from to $21 per TB on pros. 2. Compare prices for some sample configurations with the same amount of usable storage.
A drop from $35/TB to $21/TB is tiny, so tiny it doesn't usually need to even be factored in. Differences in orders of magnitude are a different deal, and they do exist from one solution type to another. Compare a 256Gb SSD or even a thumb drive, to a full on NAS with a RAID 5 array with hot spares etc. That difference is a "lot".
Factoring in that I had a stack of unused drives already, my first Synology was a 12 bay. I filled it with 1tb-6tb drives and was golden since that meant I didn't need to buy drives until I needed more space(which happened, again and again 😄)
@@littlefootfeet I have a DS2419+ and a DS1821+. Both point out I use unsupported drives when I add or replace a disk, but everything works just as it did on DSM 6, before the push by them for supported drives.
I have adopted this rule now; Any HDD capacity less than 1.5x what you can get for $200 on the SSD side, is to be considered obsolete. Currently you can get 4TB SSDs for just north of $200, that means 6TB (1.5 x 4) drives are now the smallest HDD that is worth getting. 6TB HDDs are now $90, while 4TB HDDs are $60. Up to each and everyone to put their cutoffs though!
Care to expand on your logic a bit? Getting the same capacity at around 1/4 the price sounds pretty good to me, and I don't exactly need 16+tb if all I'm after is a backup for like 2-3 TB worth of space.
@@Avruthlelbh Below $200 we are talking very little money for storage space that is just that much faster. If you need 2 TB or less SSD is a no brainer, it's like 30-40 bucks difference. 4TB, yeah, there is a small argument to make since that is like double, but, slooooooow and pulls a ton of power (5W-10W vs 1W-2W)... If you buy bulk, 2TB are less cost effective than 4TB is less cost effective than 6TB. 16TB is the sweet spot in terms of price/TB.
@@wertigonyeah, 16tb drives are the sweet spot. I went with 4x 16tb Seagate exos x18 drives in my Thunderbolt flex 8 Raid enclosure, while on my backup raid enclosure I went with another 4 16tb exos x18 / ironwolf pro drives. Now I have a raid 5 with 48tb storage, and a backup raid in another room/building, plus backblaze to save me in case both systems would fail…
Select drives based on workload and never mix workloads. If you're recording surveillance 24/7 don't mix that with other data. The surveillance activity is going to wear out drives faster. Putting other data into that mix is putting that data a risk. So you might need bigger drives for surveillance and maybe smaller drives for your other stuff. Create separate arrays to separate the workloads and buy drives that make sense for each workload.
It is important to also consider rebuild and restore performance. With a larger number of drives running RAID 5 or 6 there is a significant difference in rebuild time compared to rebuilding a mirrored array. Also there is a difference yet again when you compare something like a RAID 10 to a pool of mirrors instead. In the case of RAID 5, 50, 6, 60 and 10 your backup will be of the entire array. In the event that data loss actually occurs and you must restore from backup you are talking about restoring the entire size of the array that was lost. However, when using pools instead you can elect to restore only the data from the underline RAID 1, 5 or 6 that you lost. These are all important factors to consider when selecting how you want to build out your storage array. Also I should mention that if you are going with a lot of drives it is also possible to hybrid nest the arrays in such a way as to maximize performance while also minimizing your restore from backup time and risk of inaccessible data. For example you could take two drives and place them in a RAID 0 then take two more drives and place them in a different RAID 0. Then you take those two RAID 0 arrays and Mirror them together. You create sets of 4 drives in this method that you then pool together. The obvious con to this is that you would ideally add sets of 4 drives at a time to your pool.
There´s a interesting point to consider the multiple disk use. The SATA saturation (or other conector). If you are using the 10 Gb connector and/or VMs and docker applications.You can saturate the data bus with to much IO, exceding the SATA read/write capacity, and depending of biggest caches in SSD. And surprise, more probability of failure via hard/soft or energy. With multiple smaller disks more data can be read and write. The trick is what is the size of storage that you need and what size of NAS (slots) you have (including the extenders).
As a novice, this makes my head spin. I'm thinking about a NAS to put all my video, music, and photos in one place, with backup, and if I can afford it, allow it to be cloud storage for my kids. I have 25TB of external drives, and while it's good for storage (I'm afraid of disk failure and data loss). I'm totally lost as what to do.
depending on how many disks you have for that 25tb look at en.wikipedia.org/wiki/Standard_RAID_levels for a reference of what raid level you are comfortable with when it comes to disk failure
I got myself a few almost dead HDDs from a Chia mining farm with 30k+ hours on them. I can literally hear them dying, but they hold for 3 months now. I am running TrueNAS in RAID-Z2 setup (2 parity drives, as in RAID 6) with 6 drives total. It might not work for you as it calculates total volume based on the smallest drive size, so running different size drives may result in significant volume losses.
@@210ArtemkaI have 5 drives with roughly 70k hours each that were new when I bought them. Seagate Archive HDD 8TB's, they were throwing SMART errors out of the box, and they are still doing just fine... I thought they were gonna be dead within 3 months of purchase... :)
After an almost critical incident, hot spare makes more sense to me. If not go for dual drive fail safety in synology to stay safe. Avoid buying drives from the same batch, I went as far as buying drives from different stores and different months.
I always had this instinct but it's not discussed a lot online. I tend to buy from different batches too. What's your opinion on using different brands in a RAID 1 array for further diversification? assuming same capacity, etc..
Of course I didn't see this info until after I bought my six 6TB drives from my local shop, so that's guaranteed that mine are all from the same box, sigh. On the good side they've been spinning for 5 months now with no issues whatsoever 😊
@@shadow-wulf it's not like it'll make them go kaput at once but after dealing with mission critical databases, you just become paranoid about even an error in a batch. I'm sure you're fine! *knocks on wood*
Something to consider is drive usage. You might have an 8 bay box to be used for various things. It's best to separate workloads. If you are recording surveillance cameras 24/7, a very high write, and possibly read, situation, don't mix in other data. Not only will it slow down reads and writes of non surveillance data, but the surveillance drives are going to fail sooner and put your other data at risk. It's better to create separate arrays for each. This will also allow you to mix up drive sizes. You might need bigger drives for a bunch of camera recording for 30 days or whatever, but maybe you only need smaller drives for your other data.
Great video. One tip from lil ol me. First nas I ever used I bought 4 identical drives same make, model, style type. Unintended consequence was……. Same mtf. All the drives started failing close to each other. Next nas I made sure had a mix of different brands, different styles, mix of new and used. That should spread out the failures to different times
I had a slightly different experience from what is expressed here, though I’m not casting any doubt on the validity of the information. I have a Nimbustor 4, populated with 4x8TB, for 24TB of storage. I went for 8TB because, at the time, it represented the best bang-for-buck and gave me a total capacity (24TB in RAID 5) that I was unlikely to exceed for quite a while. The NAS itself represented the best box I could justify. Being four bay, it also gave me the opportunity to spread my expenses over a longer period. When I originally set it up, it had two drives in it. It was kind of noisy but no more than I expected. When I added a third drive, the noise and amount of disc access was much greater than it had been. Recently I upgraded to 8GB of RAM and added a fourth drive. The first thing I noticed was that the overall noise is far less than it was with three drives and almost certainly lower than it was with two drives. In fact, it’s got to the point where I rarely hear it. This probably won’t be most people’s experience but it seems to me that either by luck or design, I ended up in a sweet spot. I can’t explain it but I can hazard a guess that this is the kind of setup the designers envisioned…?
Very informative videos. You give info that is very relevant without beating around the bush. You make a good point about balancing out between the cost of NAS system and the actual storage. One option for anyone starting out could be to go for a smaller storage and then buy more/bigger HDD as the data grows and HDDs become more affordable in future, while investing in a NAS with more bays for future-proofing if really necessary. That also means hopefully sticking with the NAS with more bays for a longer time, given the initial cost.
Another specific advantage, related to the advantage of simultaneous reads & writes on multiple disks, is that you can tune a RDBMS so it purposely spreads data across multiple drives and even platters to optimize access, especially for searches.
More bays: allows expansion, means you can postphone an upgrade. Clarifying your data increase is also important. Duplicate finder is alao a good way to save money here.
Gr9 vid man! Appreciate you going thru all of the various different perspectives and angles of all of this info! My plan is 6 drives, raid 6, at least 2 systems, 1 system as backup, 1 system live, large format drives, not going to be cheap, but want the redundancy of raid and mirror, allowing up to 2 drive failures at one time. Most likely just Truenas scale at this point. Subbed and liked! Keep up the great work!
I’d always lean towards getting a bigger NAS and smaller drives rather than bigger drives and smaller NAS. There’s more options in terms of backup and space options.
I've deceided to get a NAS for home use as a media server. I'm checking out a lot of your videos to make up my mind what I actualy want and need, your videos are great, very detailed and very good explanations of the usage and for whom that particular NAS would be of interest. Kepp it up....
If a drive in a RAID configuration fails, it takes longer to restore a large driver than a smallere. Thus it takes longer time before the redundancy is restored. We are talking time in the order of days. During that period another disk failure leads to full data loss
When I first got a NAS for mass storage I got a 4 bay NAS and filled it with 4TB drives, it was nearly full after 4 years and I upgraded the drives inside it with 4 10TB drives. I back up the most important data off on the NAS over the movies.
In 2004 I had 9 Seagate hard drives fail in a 2 week period and they were sequential serial numbers. When I contacted them about a possible issue with that batch they spewed out the corporate boiler plate response saying that wasn’t the case and that their hard drives were of very high quality blah blah blah. I asked for new replacements rather than refurbished ones but they wouldn’t do that either. I’ve never sold another Seagate drive since. I doubt the few thousand drives I’ve sold over the years that weren’t Seagate are missed by them but I also never recommend Seagate because of their piss poor customer service.
I think the bit you didn't really focus on when comparing two 18TB drives vs 5 smaller drives was you dropped from a RAID Mirror to a RAID Array. That is often a big performance drop. Also replacing & rebuilding a RAID 1 mirror is faster. And if you lose both, and want to pay someone to do disk plater recovery, the probability of retrieving usage data, drops significantly when they need to reconstruct the data off multiple drives. When might you lose both RAID 1 drives simultaneously? Domestic: Voltage spike/lightning. Commercial: Some idiot unscrews the wrong thing & drops all disks on the floor.
Another point to consider is when the inevitable drive failure happens... How long does it take to rebuild the array? My 12TB RAID5 array takes ~23 hours to parity check or rebuild a failed disk. The bigger the disks, the longer the rebuild. If your bought a batch of disks from the same retailer at the same time (common thing to do)... will a 2nd disk fail during the rebuild? So another tip - buy your disks from different retailers (2 from here, 2 from there kind of strategy)... hopefully you will get disks from different manufacturing lines or at least different batches to reduce the likelihood of simultaneous failures.
I've mentioned this before, but if you are shopping for Pro drives, the only way to go is enterprise (Exos, Ultrastar). Depending on region they are much cheaper and also better.
I (mostly) agree! Just to check, did you watch my video on Pro vs Ent HDDs last year (he said, arrogantly)? If not, I would love to hear your thoughts on it - ruclips.net/video/KjA1SLKvBZM/видео.html
@@nascompares I do indeed watch most of your content (including the one you mention), it is much appreciated btw. I can of course only judge based upon my own experiences and region (Scandinavia). But for the last few years at least, prices on enterprise have been consistently lower vs. pro, which means that at least in my case pro doesn't make sense. Let me be clear, I only look for largest sizes available, save the most recently released, for the same reasons you have highlighted. Example, at time of writing and what I see pretty much consistently: Red Pro 20 TB - ≈ 563€ Ironwolf Pro 20 TB - ≈ 508€ Ultrastar 20 TB - ≈ 397€ Exos 20 TB - ≈ 377€
@@the_bogeyman. Presumably they are the "civilian" version of Ultrastar, so yes. I've never found them at less than ludicrous prices though, so it's never been relevant to me.
I've had 7X 4TB Hitachi Ultrastar drives for my NAS since 2015, and still haven't had one go bad on me. I've run it in both a RAID 6 and RAID 10 with a hot spare, and in both hardware and software (WSS) RAID modes, and recently bought another drive to make it 8 and did away with the hot spare, making it the storage for my backups. Still pretty reliable, but I wanted to replace it with SSDs. I'm a believer in minimum 6 drive arrays for NAS, for both performance and redundancy.
I always tend towards the higher end of drive sizes but normally stay a step or 2 below largest capacity so I can have a better price per TB and just know that I can get bigger for cheaper in the future I prefer having expansion bays in my system open as my storage needs grow as that always have and will continue to
yeah $/GB is best around 16-18TB but also need to consider the amount of drives vs. total storage... 2x 16TB is less storage than 4x8TB then $/GB is through the roof when you only get 16TB vs 24TB on the other setup 2 vs 4 drives that is important too
Noise is really a bummer for me now. I went from 4x3TB red to 3x8TB iron wolf, and the noise is more than double, and far more annoying in character. Heat and power also way higher. For next upgrade I want 14TB plus disks, and they are all horribly loud. I want high capacity, but low noise, low speed, low energy, since I can have multi TB of NVMe cache.
@@brandon_wallace RAID 10 has some advantages over RAID 5/6 including likely rebuild time in case of drive failure for similar/same size drives. I have used RAID 5 for many many years and know that not only are the rebuilds likely to be slow with drive failure but worse is a failed rebuild which is a real issue that does occur. With the price of larger drives going down, a nice 4 drive RAID 10 is very practical.
You missed one point. Kinda. It falls in line with the performance aspect. the larger the drive the longer the rebuild takes on a failed drive. So a 4 or 6TB drive, depending on how full the prior drive was may take 3-8 hours depending on the speed of your system / controller. But a 20TB drive could very well take 24+ hours which in turn leaves your array in a vulnerable state for a second failure. BTW this is why NO ONE should be using RAID5 anymore. The happy medium for me is somewhere between 10TB-16TB drives. They offer larger storage, but with a reasonable rebuild time.
Great information and I love the presentation. You are definitely not boring, even if some would say the subject isn't the height of glamour. I bet you could be a stand up comedian too. A lot of the useful stuff I have learned about NAS technology is from your videos. Nice one, thanks!
The required amount of redundancy depends on how you plan to use the storage. If the NAS is the primary storage, then redundancy is very important. If the NAS is backing up primary storage on other computers, then RAID might be less critical for some people. The other consideration is point in time, or off-line, copies. If you have a fire, flood, theft or a ransomware attack, it might take out the whole NAS. Therefore, you need protected offline copies. That means maybe double or even more storage. How should we provide that? Does that remove the need for redundancy in the NAS?
More drives is always better. If you have one disk drive, failure of that one drive and you could loose everything. With more drives, you can run a raid array. With options for mirrors drives. Options to strip across drives for incredible speed. Or data protection using a drive for bit checking to ensure data stays intact. Just swap out the bad drive. And then the ultimate, use them all together. Speed, reliability. So many options. More is always better.
Here is how I do it: get the price divided by total capacity to get $/TB. That is the true cost of your storage. Then you can compare apple to apple on all of your drive options and pick the cheapest one. Just note that there is a trade off. The more drives you have, the more power it is going to draw and the more points of failures there are in your system.
Having not watched the video... both have their benefits and drawbacks. Benefits: More means higher throughput can be achieved and higher levels of redundancy can be gained making the setup more resilient when it comes to disk failure. Bigger means less power draw, less vibrations and less potential heat, less physical space used and more capacity. Drawbacks: More disks is more power draw, more vibrations more heat production and more physical space used. With the added redundancy comes less capacity as the redundancy means disks are there just to cover the situation where one or potentially more disks fail protecting you from data loss in those cases. Bigger disks means less options for redundancy as you have less individual disks, less theoretical throughput and often higher cost because even though the cost per GB drops the amount of GB's per disk is significantly higher. In the end it does not matter much which one you pick as long as you first take some time and think about what your goal is with the setup maximum redundancy and not to concerned about max capacity well more disks is better. Maximum capacity and not to concerned about the data's longevity less big disks is the best option. If your chosen NAS enclosure allows for more disks than you are currently using then less but bigger might also be a good option as it will allow you to grow the storage capacity over time.
I enjoyed this video very much! Very informative!!! What I would have like to see is a graph that shows where the flipping point is to decide on more or larger drives, including the NAS itself.
I have an 8 bay DS1821 with 4x 8TB. The rationalisation was that as my data needs grew I could either add more 8TB or on the basis that over time costs would come down start adding say 12TBs and through SHR move to an all 12TB RAID. So far so good.
I often recommend this very thing for those who have their heart set on a 4 bay but have a little more budget left (or have over specced their day 1 storage needs significantly - slice a TB or 2 off each drive, pump that money into more bays). There are arguments on both sides (i.e drive availability down the line), but if you are running on an SHR, then that's less pertinent.
I just built my own NAS after going through a 2-bay NAS and then added a 4-bay NAS. I built one with 18 HDs and it is much more expandable. I am using UNRAID and it has been great so far. It is much faster and I have so much more capacity. When I need more, I will replace some of my 6 TB drives for 12 TB drives.
I just bought WD Gold 20TB. Crazy fast Built-in NAND and transfer speed up to 500MB/s !!! and much less noise !! I always love larger drive than small one. Because you will save money to throw away small HDD.
The answer is simple: the best is to have lots of big drives! Crib the storage perspective, of course, not the noise/power consumption. Of course, with larger drives one should be very sure of the backups. And preferably use 2-disk redundancy to boot. It may be also result in higher ram usage.
Actually, I think in most case It depends on how many IO is on your motherboard such as, if you have 4 SATA ports and you used 2, then you will soon realize if you buy too many low-volume hard disks, you will soon be out of IO to let you upgrade in the future
Still not clear which is best 🤔🤔, so I bought a QNAP TS-1232PXU-RP-4G 12bay ($1475) and put in 12 x 20TB drives (Ironwolf Pro $329 each). After RAID 6 have 185TB available 😉😉2 x 10gbe, 2 x 2.5gbe network connection. Bought 13 drives to have a spare. Total cost before tax: $5752.
OK Typed up a long post and then lost it all... So here is a short version. Check the manufacturers HCL (hardware compatibility list) before buying drives. And if you don't find a drive of the capacity you want then consider if it's worth the risk. RAID controllers can be real finicky about drives. You may feel SATA and SAS is mature tech and there should be no compatibility problems, but there are and there will be more. I've worked with (from memory) Adaptec, Areca, Raidcore, 3Ware and LSI. Sometimes the compatibility problems are blatantly obvious, but sometimes they are a creeping problem that takes time to develop, and they don't get better with time. Sometimes a firmware upgrade of the drives or the controller can help, but there's no guarantee that either is coming if you start out with incompatible hardware. Also stress test the arrays before your start using them. Run every storage test you can think of on them, and then try some more. Check the RAID logs and take note of any warnings. You don't want warnings! Not even the non critical kind. Make sure there's as little vibrations as possible. Vibrations can play havoc with RAID arrays even if they are not strong enough to cause a head crash. Also don't use Shingled magnetic hard drives. They are a pain when used for RAID. Temperature! A interesting paper published by a storage company probably a decade ago showed that the ideal running temp for HDD's seems to be between 35 and 45 °C. Higher or lower temperatures showed increased failure rate. But don't take this as gospel. However we do know that high temperatures are bad in general, and 40°C is a quite easy target for HDD's.
I feel like power consumption… isn’t a factor. You’re going to consume more power. You should expect that. Also noise… isn’t a factor. HDDs make noise. If you don’t want noise, don’t get HDDs. Here are my take-aways: 1. Don’t just get 2 drives. Because you’ll end up using one for parity only and waste the space you could use. 2. Get 4-5 smaller drives at once so you can benefit from the performance boost. This also ensures you can have more useable space over all. Unless you have the cash to fill out 4-5 18TB drives in your NAS, just get smaller drives. Then you can have better performance and more redundancy. If you didn’t want high power consumption and lots of noise, you shouldn’t be buying a NAS and filling it full of HDDs.
Bigger drives are better. But if the data is important to you, the cheapest way (if it is not for professional needs), get the biggest drive vs price you need, and have an extra one as backup you dont use except for backup. Keep that backup away from power in some storage shelves or so. Hdd you dont use last very long. Had a drive from 10 years ago that I almost never used and put it in 'cold' storage, so unplugged in a shelves, and worked like a Sharm.
I've been toying with getting an expansion unit for my 1019+ to get more drives or upgrading the current ones I have. 2 - 8 TB, 3 - 4 TB, 2 - 500 GB cache.
Thank you for the thorough video! Perhaps one other aspect to consider is the best way to increase space with a 4-bay NAS that has expansion capability. I have a 918+ with 4x4TB. Shall I replace the drives with 4x8TB drives, or shall I daisy-chain a second 4-bay NAS with another array of 4x4TB?
I think I’d go for replacing the 4x4 TB with 4x8 TB because, well otherwise you need to buy an expansion unit and additional drives. You then can do that again. But at last you have to replace all 9 drives.
*what he said* - Yeah, buy the new larger drives (8TBs are regularly on offer too). Plus, if you can stretch to it, buy a really low end 4 Bay and then set it up as a remote backup device. A proper value one with a Realtek CPU. Then have it on the network or remotely at a friend/colleague/family member's home. Sync them... boom... BACKUP CITY!
I'm no Synology expert and these types of questions can be best placed upon NASCompares forum. For what it is worth; we own a lot of NAS and never ever expand a NAS, we either replace the NAS (or HDD's) or place another NAS into the network. The problem with expanding your NAS is that there is often a catch; either you lose (some?) redundancy, or lose speed & performance. Plus the risks when something happens between either of them, how will it cope with the problem, will it not destroy/crash the volume? And the tipping point is often more the costs; what is the budget/how much are your plans going to cost. Expansion-units are quite expensive, and with the (technical) drawbacks, often the better choice is looking into alternative paths to increasing the capacity..
One idea to consider is the difference in power for say 4 big drives (say 14TB each with 1 parity) versus 12 small drives (say 4TB each with 2 parity). The comparison of a small size storage ( smaller than 20 TB ) I think was the focus of the video and not larger size storage (35+ TB storage sizes. Overall the video hit the points I think are important.
you start with saying smaller drives are cheaper, and while they are cheaper as singles, if I were to buy a skyhawk 4tb its 21.5 per tb, a 20tb is 17.2 per tb. an exos is bigger disparity, in favor of larger drives.
Also remember: an active raid is not a permanent backup solution, it's a stop-gap. You should always do regular backups to an offline media as well. I suggest a raid 5/6 for active use and backups then a mirrored external for offline.
And when backing up remember this law in computer backups “one=none, two=one” the more backups the better. Always good to have offsite backup, onsite backup, and a locked in a bank vault backup.
@@Technichian462 exactly. I try to put my important documents on as many drives as I can. I think my most important documents are backed up over 5 times lol. I even use my old phones as storage. You never know when it'll save you.
@@Technichian462 one question we were asked in our information technology class was "now what do you do if your tax returns are backed up three times in a house that burns down and you can't get access to either of them?"
@@davel4030 Thts why backups are always done/kept off site. You need to weigh the importance of your data. Critical mission data, data that you cant afford to lose, gets back up every day, maybe even every hour, and it gets backed up sequentially. Indeed, that means you can have hundreds of backups. You label them by date and time. That back up becomes a time machine for your data. Allows you to wipe the system clean, and reload it he system from last week, yesterday, even last months data. If you are ever infected with malware you will wish you had that sort of back up. I’ve seen it happen. Lots of important data that could never be replaced lost forever. Millions of dollars worth of data just gone. Never to be replaced. The company went bankrupt. I dont think the owner ever did anything else after. All he needed to do was spend a few thousand dollars, to have multiple raid arrays running and offsite back up of that. With backup running every night, with sequential saves going every days for two weeks. He could have been like Elon Musk, but not like a shyster p.o.s.
Appreciate the mention that if you buy a number of the same drive from the same vendor you are likely to get drives from the same lot which if that lot has a problem means your risk for trouble is increased.
I use 8 x 8tb as standard in my 8 bay nas's, JBOD. Reason is, I use a 2nd and 3rd nas as backups, and if a drive fails I just copy the data onto its replacement, that way I just keep 1 nas running, otherwise 2 would be on all the time, I do a backup using goodsync once or twice a month, I tried synolgys drive sync, too automatic. I like the control of goodsync.
Honestly I just have a bunch of 4tb red drives (every seagate I have ever bought failed within a year.. all of them..) My reasoning is its easier to replace a 4tb drive than an 18tb drive. Both in cost and in time. If I fail on a rebuild then I only loose 4tb of data. (I use unraid) I only backup what I can't get back (pictures, home movies, etc. I can always re rip my dvds and such. 3-2-1 can get expensive otherwise. Especially with larger drives. The little nas boxes seem pretty neat but frankly an old pc with an hba card is all you need. Buy unraid once (or use truenas, openmediavault, linux, whatever you prefer) works. I prefer unraid because of the way it works. Even if you fail on a rebuild you only loose whats on the failed drive. With raid you loose the whole pool. With nas boxes your upgrade path is kinda expensive. With other options you can just use your old pc when you upgrade. My 2 cents worth. A lot of options. Depends on risk, time, and finances. Everyone's mileage will vary.
Having just replaced or added HDD's to both my 4 bay Synology and my 4 bay QNAP, my next NAS purchase to replace one of these will be an 8 bay. And I'll install 3 larger HDD's in RAID 5 or Synology Hybrid RAID and add new and larger drives as the need arises and prices fall. The worst thing about smaller NAS is the cost of pulling out perfect good, decent sized HDD's because you need to replace them with larger drives. This is especially problematic with QNAP as you need to have all drives be the same size. With Synology's SHR, drive sizes can be mixed ( although even that is not perfect). The extra cost of an 8 bay vs a 4 bay is easily saved by the ability to continue to use my older, smaller HDD's. Don't be deterred by up front costs; look at longer term use and future purchases and redundancies.
Cheaper more drives but what about power consumption? More watts consumed or is the same? Let’s say will last 6 years and had to pay more electricity ⚡️ during those 6 years that also impact
The "not all eggs in one basket" is a bad figure for RAID5. Speaking from probability of losing your data, using LESS drives is better. Let's say the drive failure rate is 3% per year. - 2x6 TB RAID1: probability of losing ALL data (2 drives fail same year) = 0.03 * 0.03 = 0.09% per year - 4x2 TB RAID5: probability of losing ALL data (2 drives failure plus 3 drives failure plus 4 drives) = 0.518 % per year so the RAID1 is A LOT safer BECAUSE it's using a double safe basket instead of multiple baskets that are connected and ALL fail if 2+ fail
Or, Dell T630 LFF server £350. Used 2GB cache RAID controller £110 6x used 8TB HC520 drives, £70 each. Needed a server as well as a lot of storage, so leaving out CPU & SSD upgrades.
Per gig is cheap now-a-days. However, i don't think i should admit i still use Barracuda desktop drives in my NAS... Their cheap, compared to RED drives and IronWolf.. Besides,from past experiences, they 'whine' allot in idle mode... Could of just been bad drive, but i doubt. These were 4TB drives Also, power-saving can make up the difference between buying big drives... The presumption your making is NAS's are designed to be on all the time and active all the time, which is not always true. There is always going sections of 'idle', time, (particularly after midnight),. If you have Scheduled backups going on a QNAP, your gonna allow a few hour either way before the next starts to prevent possible increased failure. In that time space, the dives will spin down after 30 mins (usually) thus saving power. If you work that over a given year, that's still a bit of energy saved right there.
huge enterprise drives can go 300mb/s sequential while a 4 or 6tb drive usually cant even hit 200mb/s especially if they're 5400rpm so fewer disks can be as fast depending on the size difference. Also, the power use is substantial when using more disks. disk power usage can be more than the rest of the entire system combined when talking about 10+ disks
I've got this idea recently.. In theory, what about using classic non-NAS type 2,5" SATA SSD's? For example four 2TB SSD's in RAID 5? I know, it's more expensive. But it's silent, less power consumption and if I would have money for it? Just for home using, not really too much transfering data and important data will be backed up to external drive too. Is it there some statistics or something? What's yours opinions guys? I have few data SSD's in pc for long long time using by daily basis and they are immortal.
This is what i am pondering aswell, fast archive can be well maintend , duplicates removal etc. Some external snapshots for less intresting data etc. Question is for the installed OS drive mirrors, put that to an expensive storage? How many snapahots to keep etc.
Nice video! I always calculate the price per TB. ;-) As that way there is often a HD-size (and above) that the price per TB is higher than the smaller/previous-size one. (price per TB) We often opt for a wee bit smaller but more drives. As we always have 2x cold standby HD's (thus unused!) near the NAS. But more drives also means more power-use as you already indicated and also more heat. (more cooling might be needed) Business-wide you should replace your HD's every 3 years (5 years, max) but for the average user at home, you replace when really needed. (or after 10 years?) To reduce risks, we also opt for multiple NAS. As a NAS may fail (power-supply, firmware-upgrade, bricking) With larger drives the rebuilding of the RAID also may take longer., when something does go wrong. BTW, also worth noting, the weight of the NAS might become significantly higher when you are using more drives (noticeable after 8x drives IMHO) Generally speaking, the 6TB and lower are robusts as rocks (longevity), 8TB to 10TB are often the sweet-spot for pricing in my experiences. Word of advice: buy as many drives as can fit in your NAS as down the line by the time you want to buy additional drives, the manufacturer may have moved-on to newer models..
There are indeed so many arguments and I do not want to use the words pro's and con's but rather benefits and caveats. Your video is quite clear there is no definite answer. Plus, I would say, personal preferences. But excellent to line-out some of things one can think about when making such choices. As a famous Dutch soccer player once said: every disadvantage has its advantages ...
As always, wonderful observations man. This video was largely born of an issue I faced with a user in the advice section that was buying a 24 bay and wanted to partially populate. I did highlight (as you did) that all too often smaller or more niche drives can go 'end of life' or changed (ANYONE WHO BOUGHT FIRST GEN WD GOLD HDDS - YOU KNOW!), but it wasn't especially heeded. 15 mins into composing my email to the chap and I realized it was effectively an entire video script. The same is true of a UPS video that should go out in the next week or so (though that is even more niche and the pros and cons even more ambiguous). Also, I know this vid is only like 2 hours old, but I'm surprised my points about HDD noise are not getting highlighted more. Sometimes I genuinely feel like I am the only sod who is talking about how bloody noisy ent/pro drives are! Too, TOO many phone/video editors are buying 4 bays with 10GbE, stuffing it with 20TBs and then getting a shock that it sounds like R2D2 having a panic attack! I'm quite looking forward to seeing your thoughts on my video coming soon (genuinely called) '10 Dumb Questions about NAS that are not actually dumb' (part 1 in possibly a 3 part series). Have a great weekend bud.
@@nascompares You most certainly have a very valid point about the noise, that's also why I have not all of our NASses turned-on all the time. (and some experiences with WOL & 10GBe haha) And even then, when I do a full shutdown (once so often), the true silence can be deafening! Although modern-day high throughput switch's fans can, especially when coming of age, become a much larger nuisance in my experience. That's also why I have the truly sound-dampened 19 inch racks (Ucoustic Edge) in use. I always look forward to your videos, as they make me think, consider and sometimes even reconsider. And think outside the (padded) box, being more creative, flexible, and sometimes be amazed. (in the positive sense, of course!) To me personally the dumbest question is the question not being asked, if you pardon the pun. Thank you again for this gem of a video! You too have a great weekend!
It depends. If your nas it is just for fileserver and a few clients go foe bigger disk. If youf nas contains db's and lot of clients, better more disk..
I would have divided the presentation between Consumer needs/suggestions and business needs/suggestions: in case of business you need IOPs and you do not care about noise because you have a datacenter (for small that it can be), so smaller disks in higher number give you better performance anche noise means Enterprse disks which have higher number of TB/year. at the same time as enterprise, for small that you can be, you should implement the 3-2-1 backup rules or at least having a backup site. the additional point is having a redundant power supply on your NAS. for the remaining part, what you said is perfect for a consumer site.
The issue/barrier I will always have with tackling any subject with separate Consumer/business POVs, is that it immediately overcomplicates the subject. That's why I make separate videos for home/business. I recommend you watch my Pro Vs Ent drive series
Thanks mate for the great info. I have a 5 yr. old TS-653A with 5 WD60EFRX and I'm stuck with 1gbps connectivity. I'd like to get at least 2.5gbps upgrade so I'm going to get another NAS with faster networking and use it as a staging NAS. I'll add a 6th drive to my 653A and move it to my in-laws house and work out syncing between the 2 NAS.
It depends on what you need... More I/O - IOPS vs. need for more space. You can't really get both... Great video! I have done a lot of videos on this very topic...
You talk about the possible more possible points of failure.. but you miss the big point with Raid 5 vs Raid 1... raid 1, if one drive dies.. you take it out.. order a new one.. re-raid it when you get it.. with raid 5.. when one fails.. you have to get another drive and rebuild it.. b ut while that is going on.. the entire raid is OFFLINE.. so I normally recommend.. if you RAID 5.. order a spare to minimize the downtime...
If you have two large capacity drives and one of them fails…that’s a HUGE SLOG to replace in one hit. If you have lots of small drives and one fails it’s not going to hit you as hard when you suddenly have to get the replacement.
Why not "more bigger hard drives?" When I bought my NAS I got four of the biggest HDD I could at the time. Now they make drives that are almost twice as large, so I'm jonesing for an unneeded upgrade. The noise is a legitimate concern, but I think for most people the best solution is to try and have a wiring closet / server room type setup and put it in there where you won't really hear it. Of course, the problem with that is that the house basically needs wired for ethernet with a patch panel. I wish I had that.
This isn’t strictly relevant to this specific video but I’m asking advice I’ve been given the task of assembling a moderately large NAS for a small company. I’ve decided I am going to include cache but it’s the type to get I’m confused by. It happens to be a Synology NAS I’ve gone for, and I noticed that specific types of M.2 Sara or nvme are recommended. It basically narrowed it down to WD Red, FireCuda 520 and Synolgy’s own 400 or 800Gb Nvme. My initial reflex was that it was probably a good idea to go for Synology as it’s the same make as the enclosure but 400Gb of Synology SNV3410 Cache is about twice as expensive as 1Tb of WD Red nvme. Why is this and is there anything that justifies this price difference?
I prefer drives under 4TB as I find them to be more bulletproof. Bigger video and game files as well as AI models have caused me to now need large capacity drives. I been on a computer since DOS. My first PC had a 100MB IDE drive. I used that drive till IDE was fazed out and SATA became the standard. It never died nor slowed down. But it did sound like the Predator from the movie. I stopped using it cause IDE was just too slow. I am 40 years old and I have drives that I had when I was a teenager that still work fine today. The one that runs my pfSense is a 2.5 inch that has been in a laptop for about 6 years till it went in my Desktop for extra storage for about another 5 years and now runs my firewall. Its the only old drive I have that clicks. Been clicking for years now but will not die. But every singe 4TB or larger drive I get will need to be replaced at some point cause they are sensitive like lil girls. Vibration or noise or impact or temps or looking at it too long will break it. I have spare large drives just in my closet. No small drives cause they just won't stop. That 2.5 120GB has been in bumpy cars, dropped 100's of times, bumped into, ran sitting upright, ran upside down, sideways, slanted, its older than some peoples children, and still clicks along. Its seen soo many video drivers, windows updates and PornHub. But 4TB and up...a loud noise might startle the thing and make it slap its forehead with the back of its hand as it faints. You gotta wake it back up in the controller. Shaking a grown man and he will most likely survive. Don't do that with a baby. But HDD's is different. You can shake the baby HDD's but if you shake the big grown HDD's they are dead dead dead. We are at a age now where files are big now so I am building out a server rack using large HDD's not because I need a server rack but because the server rack needs a safe place to even be a server or NAS like Hollywood. PC's are built different. They will save mkv's and load steam in the streets of Brooklyn even after being dropped violently because a bee flew in your face and swatting at it and missing the bee made you smack the PC on to the ground. Them small capacity drives are built like 50 Cent
It would be really interesting to me if you could make videos at regular intervals for things like When will the price of hard drive come down and when will we see bigger capacities in the small classes.
That's a good idea, but would be quite a lot of videos with quite short shelf lives. That said, I really, REALLY recommend the Price per TB tool that Eddie built (nascompares.com/2022/12/28/how-to-choose-the-best-value-hard-drive-and-best-price-per-tb-get-it-right-first-time/) This will allow you to enter the parameters of drives, capacity, brand, RAID and more - then it spits out drives (priced as of 'now') on amazon. Its updated daily and if you are looking for the best way to track how much TBs your money can buy you in 2023, its the best there is tbh. Hope this helps!
currently my Media collection is currently under 3TB so I do not need a very big set up yet. Currently using 3 2TB reds in a raid 5. About to upgrade to 3 3TB drives in a raid 5 Then have a single 10TB HDD as a Back up.
I dont think we should only look at size vs number but also failure rate. I would rather go with a disk that doesn't fail on me that often then the one i need to buy a new disk every few years and rebuild the RAID.
If your Data is very very very important than use a 4 bay or larger, with larger Harddrives in Raid 1. in a 4 bay for example you have 3 Discs of failure.
If you have seriously important data don't use a NAS use a SAN, lots of drives only generally gives a performance increase only for reading not writing, RAID on NAS devices usually has some restrictions based on implementation of the standards of the supplier vs on board raid provided by server manufacturers. NAS providers are great at vendor lock-in. Also make sure you buy drives compatible with the NAS as they don't cover warranty issues otherwise. Generally pro series drives offer 5year warranty non pro are 2-3 years
You know what, I have been researching to having a solution for iCloud and Google One Drive because I am really struggling saving my image RAW files and videos. Then all on a sudden a photographer Tony Northrup brought the light of a NAS! I did not know what NAS is until couple of weeks ago! Then I started to do my own research and found you. I know you dont have smooth voice and attraction catching vocal gestures, but I find myself in you, I would want to express my research so that people can decide what's best for them. I have found the same agony in you. You are like a tech big brother who wants to advice whats best for us instead biased brand marketing. I like your videos. Just wanted to pay my gratitude because I know, a small wish can boost up the moral energy a lot cause you have done so much research, night and day sleepless time. I know for the video but I know it's for the people whom you want to help so desperately. Thank you so so much.
The time to rebuild an array with 1 of several smaller failed drives verses the time it takes to rebuild an array with 1 of two large drives is important to me as a home user.
I think it really depends where you are and what you can get. In some countries, it's cheaper to get bigger drives and a two bay NAS, where other places, smaller hard drives and a four bay NAS ended up being the better option. I've always tackled it as a your-mileage-may-vary scenario.
Although if you want maximum capacity, a bigger NAS is the way to go.
"I think it really depends" could easily be the sub title of this video tbh. All of the points I touch on here are general and "more often than not" categories. I guess the main thrust of it is that there is no definitive answer to Big drive v Many drives, but there are enough smaller points to make one argument more compelling than the other for different users. Personally...I carry 87 USBs in a carrier bag and I shan't have it any other way....
Add price to this whole inflation, cost of living everything going up, it basically limits people ..
"You'd like to get a large drive, if you could only afford it"
@@Tech-geeky Well, prices being what they are now, you'll really need to have some abnormally large amount of data for some reason, to get to where you're economically limited realistically. Second hand HDD's cost very little, have enough of them on hot-spare and cold-spare and you can run your array on a potato PC from 10yrs ago, that you have a pair of mirroring each other. Boom, Bob's your uncle - nearly nukeproof. Heck, if you want to be nukeproof, get a cheap laser printer and print the data out into a fire proof bin. Most things can be done on very limited budgets with time and effort, which you ought to have if your budget is severely limited. In most cases people have at least a lot of either time or money, both can solve these sorts of things, just different ways.
Really interesting video, thanks. Here are a couple of ideas on presentation of the numbers: 1. Produce a line graph of price against drive size. That would make it clear that the price per terabyte drops a lot from 1TB to 20TB drives - $35 from to $21 per TB on pros. 2. Compare prices for some sample configurations with the same amount of usable storage.
A drop from $35/TB to $21/TB is tiny, so tiny it doesn't usually need to even be factored in. Differences in orders of magnitude are a different deal, and they do exist from one solution type to another. Compare a 256Gb SSD or even a thumb drive, to a full on NAS with a RAID 5 array with hot spares etc. That difference is a "lot".
Factoring in that I had a stack of unused drives already, my first Synology was a 12 bay. I filled it with 1tb-6tb drives and was golden since that meant I didn't need to buy drives until I needed more space(which happened, again and again 😄)
Did it complain about them not being “synology compatible” drives?
@@littlefootfeet I have a DS2419+ and a DS1821+. Both point out I use unsupported drives when I add or replace a disk, but everything works just as it did on DSM 6, before the push by them for supported drives.
I have adopted this rule now; Any HDD capacity less than 1.5x what you can get for $200 on the SSD side, is to be considered obsolete.
Currently you can get 4TB SSDs for just north of $200, that means 6TB (1.5 x 4) drives are now the smallest HDD that is worth getting. 6TB HDDs are now $90, while 4TB HDDs are $60. Up to each and everyone to put their cutoffs though!
Care to expand on your logic a bit? Getting the same capacity at around 1/4 the price sounds pretty good to me, and I don't exactly need 16+tb if all I'm after is a backup for like 2-3 TB worth of space.
@@Avruthlelbh Below $200 we are talking very little money for storage space that is just that much faster. If you need 2 TB or less SSD is a no brainer, it's like 30-40 bucks difference. 4TB, yeah, there is a small argument to make since that is like double, but, slooooooow and pulls a ton of power (5W-10W vs 1W-2W)...
If you buy bulk, 2TB are less cost effective than 4TB is less cost effective than 6TB. 16TB is the sweet spot in terms of price/TB.
@@wertigonyeah, 16tb drives are the sweet spot. I went with 4x 16tb Seagate exos x18 drives in my Thunderbolt flex 8 Raid enclosure, while on my backup raid enclosure I went with another 4 16tb exos x18 / ironwolf pro drives. Now I have a raid 5 with 48tb storage, and a backup raid in another room/building, plus backblaze to save me in case both systems would fail…
Select drives based on workload and never mix workloads.
If you're recording surveillance 24/7 don't mix that with other data. The surveillance activity is going to wear out drives faster. Putting other data into that mix is putting that data a risk.
So you might need bigger drives for surveillance and maybe smaller drives for your other stuff. Create separate arrays to separate the workloads and buy drives that make sense for each workload.
It is important to also consider rebuild and restore performance. With a larger number of drives running RAID 5 or 6 there is a significant difference in rebuild time compared to rebuilding a mirrored array. Also there is a difference yet again when you compare something like a RAID 10 to a pool of mirrors instead. In the case of RAID 5, 50, 6, 60 and 10 your backup will be of the entire array. In the event that data loss actually occurs and you must restore from backup you are talking about restoring the entire size of the array that was lost. However, when using pools instead you can elect to restore only the data from the underline RAID 1, 5 or 6 that you lost. These are all important factors to consider when selecting how you want to build out your storage array.
Also I should mention that if you are going with a lot of drives it is also possible to hybrid nest the arrays in such a way as to maximize performance while also minimizing your restore from backup time and risk of inaccessible data. For example you could take two drives and place them in a RAID 0 then take two more drives and place them in a different RAID 0. Then you take those two RAID 0 arrays and Mirror them together. You create sets of 4 drives in this method that you then pool together. The obvious con to this is that you would ideally add sets of 4 drives at a time to your pool.
There´s a interesting point to consider the multiple disk use. The SATA saturation (or other conector). If you are using the 10 Gb connector and/or VMs and docker applications.You can saturate the data bus with to much IO, exceding the SATA read/write capacity, and depending of biggest caches in SSD. And surprise, more probability of failure via hard/soft or energy. With multiple smaller disks more data can be read and write. The trick is what is the size of storage that you need and what size of NAS (slots) you have (including the extenders).
As a novice, this makes my head spin. I'm thinking about a NAS to put all my video, music, and photos in one place, with backup, and if I can afford it, allow it to be cloud storage for my kids. I have 25TB of external drives, and while it's good for storage (I'm afraid of disk failure and data loss). I'm totally lost as what to do.
shr1 and 1 external hdd backup
depending on how many disks you have for that 25tb look at en.wikipedia.org/wiki/Standard_RAID_levels for a reference of what raid level you are comfortable with when it comes to disk failure
I got myself a few almost dead HDDs from a Chia mining farm with 30k+ hours on them. I can literally hear them dying, but they hold for 3 months now. I am running TrueNAS in RAID-Z2 setup (2 parity drives, as in RAID 6) with 6 drives total. It might not work for you as it calculates total volume based on the smallest drive size, so running different size drives may result in significant volume losses.
😮😮😮😢😢😢😢😢😢😢😢😢😮😮😮😮😢😮😮😮😮😮😮😢
@@210ArtemkaI have 5 drives with roughly 70k hours each that were new when I bought them. Seagate Archive HDD 8TB's, they were throwing SMART errors out of the box, and they are still doing just fine... I thought they were gonna be dead within 3 months of purchase... :)
After an almost critical incident, hot spare makes more sense to me. If not go for dual drive fail safety in synology to stay safe. Avoid buying drives from the same batch, I went as far as buying drives from different stores and different months.
I always had this instinct but it's not discussed a lot online. I tend to buy from different batches too. What's your opinion on using different brands in a RAID 1 array for further diversification? assuming same capacity, etc..
Of course I didn't see this info until after I bought my six 6TB drives from my local shop, so that's guaranteed that mine are all from the same box, sigh. On the good side they've been spinning for 5 months now with no issues whatsoever
😊
@@shadow-wulf it's not like it'll make them go kaput at once but after dealing with mission critical databases, you just become paranoid about even an error in a batch. I'm sure you're fine! *knocks on wood*
Hot spare means if you don't have a failure for a long while by the time you do your hot spare is out of warranty.
Something to consider is drive usage. You might have an 8 bay box to be used for various things. It's best to separate workloads. If you are recording surveillance cameras 24/7, a very high write, and possibly read, situation, don't mix in other data. Not only will it slow down reads and writes of non surveillance data, but the surveillance drives are going to fail sooner and put your other data at risk. It's better to create separate arrays for each. This will also allow you to mix up drive sizes. You might need bigger drives for a bunch of camera recording for 30 days or whatever, but maybe you only need smaller drives for your other data.
Great video. One tip from lil ol me. First nas I ever used I bought 4 identical drives same make, model, style type. Unintended consequence was……. Same mtf. All the drives started failing close to each other. Next nas I made sure had a mix of different brands, different styles, mix of new and used. That should spread out the failures to different times
I had a slightly different experience from what is expressed here, though I’m not casting any doubt on the validity of the information. I have a Nimbustor 4, populated with 4x8TB, for 24TB of storage.
I went for 8TB because, at the time, it represented the best bang-for-buck and gave me a total capacity (24TB in RAID 5) that I was unlikely to exceed for quite a while.
The NAS itself represented the best box I could justify. Being four bay, it also gave me the opportunity to spread my expenses over a longer period.
When I originally set it up, it had two drives in it. It was kind of noisy but no more than I expected. When I added a third drive, the noise and amount of disc access was much greater than it had been.
Recently I upgraded to 8GB of RAM and added a fourth drive. The first thing I noticed was that the overall noise is far less than it was with three drives and almost certainly lower than it was with two drives. In fact, it’s got to the point where I rarely hear it.
This probably won’t be most people’s experience but it seems to me that either by luck or design, I ended up in a sweet spot. I can’t explain it but I can hazard a guess that this is the kind of setup the designers envisioned…?
"I hate seagulls!" While pointing up was too funny
Very informative videos. You give info that is very relevant without beating around the bush. You make a good point about balancing out between the cost of NAS system and the actual storage. One option for anyone starting out could be to go for a smaller storage and then buy more/bigger HDD as the data grows and HDDs become more affordable in future, while investing in a NAS with more bays for future-proofing if really necessary. That also means hopefully sticking with the NAS with more bays for a longer time, given the initial cost.
Another specific advantage, related to the advantage of simultaneous reads & writes on multiple disks, is that you can tune a RDBMS so it purposely spreads data across multiple drives and even platters to optimize access, especially for searches.
More bays: allows expansion, means you can postphone an upgrade. Clarifying your data increase is also important. Duplicate finder is alao a good way to save money here.
Gr9 vid man! Appreciate you going thru all of the various different perspectives and angles of all of this info! My plan is 6 drives, raid 6, at least 2 systems, 1 system as backup, 1 system live, large format drives, not going to be cheap, but want the redundancy of raid and mirror, allowing up to 2 drive failures at one time. Most likely just Truenas scale at this point. Subbed and liked! Keep up the great work!
I just received two new WD Red drives from Amazon. Each had ENTIRELY different packaging and different stock numbers. Same size drives.
I’d always lean towards getting a bigger NAS and smaller drives rather than bigger drives and smaller NAS. There’s more options in terms of backup and space options.
I've deceided to get a NAS for home use as a media server. I'm checking out a lot of your videos to make up my mind what I actualy want and need, your videos are great, very detailed and very good explanations of the usage and for whom that particular NAS would be of interest. Kepp it up....
If a drive in a RAID configuration fails, it takes longer to restore a large driver than a smallere. Thus it takes longer time before the redundancy is restored.
We are talking time in the order of days. During that period another disk failure leads to full data loss
THAT is a bloody good point!!! Cheers for spotting the omission. Have a great weekend man
When I first got a NAS for mass storage I got a 4 bay NAS and filled it with 4TB drives, it was nearly full after 4 years and I upgraded the drives inside it with 4 10TB drives.
I back up the most important data off on the NAS over the movies.
In 2004 I had 9 Seagate hard drives fail in a 2 week period and they were sequential serial numbers. When I contacted them about a possible issue with that batch they spewed out the corporate boiler plate response saying that wasn’t the case and that their hard drives were of very high quality blah blah blah. I asked for new replacements rather than refurbished ones but they wouldn’t do that either. I’ve never sold another Seagate drive since. I doubt the few thousand drives I’ve sold over the years that weren’t Seagate are missed by them but I also never recommend Seagate because of their piss poor customer service.
I think the bit you didn't really focus on when comparing two 18TB drives vs 5 smaller drives was you dropped from a RAID Mirror to a RAID Array. That is often a big performance drop. Also replacing & rebuilding a RAID 1 mirror is faster. And if you lose both, and want to pay someone to do disk plater recovery, the probability of retrieving usage data, drops significantly when they need to reconstruct the data off multiple drives.
When might you lose both RAID 1 drives simultaneously? Domestic: Voltage spike/lightning. Commercial: Some idiot unscrews the wrong thing & drops all disks on the floor.
Only mirrors are in my opinion safe. Too many stories about crashs during RAID array builds.
Another point to consider is when the inevitable drive failure happens...
How long does it take to rebuild the array?
My 12TB RAID5 array takes ~23 hours to parity check or rebuild a failed disk.
The bigger the disks, the longer the rebuild. If your bought a batch of disks from the same retailer at the same time (common thing to do)... will a 2nd disk fail during the rebuild?
So another tip - buy your disks from different retailers (2 from here, 2 from there kind of strategy)... hopefully you will get disks from different manufacturing lines or at least different batches to reduce the likelihood of simultaneous failures.
I've mentioned this before, but if you are shopping for Pro drives, the only way to go is enterprise (Exos, Ultrastar). Depending on region they are much cheaper and also better.
I (mostly) agree! Just to check, did you watch my video on Pro vs Ent HDDs last year (he said, arrogantly)? If not, I would love to hear your thoughts on it - ruclips.net/video/KjA1SLKvBZM/видео.html
@@nascompares I do indeed watch most of your content (including the one you mention), it is much appreciated btw.
I can of course only judge based upon my own experiences and region (Scandinavia). But for the last few years at least, prices on enterprise have been consistently lower vs. pro, which means that at least in my case pro doesn't make sense.
Let me be clear, I only look for largest sizes available, save the most recently released, for the same reasons you have highlighted.
Example, at time of writing and what I see pretty much consistently:
Red Pro 20 TB - ≈ 563€
Ironwolf Pro 20 TB - ≈ 508€
Ultrastar 20 TB - ≈ 397€
Exos 20 TB - ≈ 377€
Are WD gold as good as Ultrastar?
@@the_bogeyman. Presumably they are the "civilian" version of Ultrastar, so yes. I've never found them at less than ludicrous prices though, so it's never been relevant to me.
@@blcjck8121 for some reason, wd gold 16tb are 100€ more expensive than wd ultrastar 16tb here. I am filling my ds920+ with ultrastars.
I've had 7X 4TB Hitachi Ultrastar drives for my NAS since 2015, and still haven't had one go bad on me. I've run it in both a RAID 6 and RAID 10 with a hot spare, and in both hardware and software (WSS) RAID modes, and recently bought another drive to make it 8 and did away with the hot spare, making it the storage for my backups. Still pretty reliable, but I wanted to replace it with SSDs. I'm a believer in minimum 6 drive arrays for NAS, for both performance and redundancy.
At this point, I've settled on the EXOS drives going forward for the capacity, cache, price, and warranty.
I always tend towards the higher end of drive sizes but normally stay a step or 2 below largest capacity so I can have a better price per TB and just know that I can get bigger for cheaper in the future I prefer having expansion bays in my system open as my storage needs grow as that always have and will continue to
Thank you for this video sir! Good as always
Toshiba 16TB drives have been cheap for some time. Now the weet spot seems 18TB, but ymmv
yeah $/GB is best around 16-18TB but also need to consider the amount of drives vs. total storage... 2x 16TB is less storage than 4x8TB then $/GB is through the roof when you only get 16TB vs 24TB on the other setup 2 vs 4 drives that is important too
Noise is really a bummer for me now. I went from 4x3TB red to 3x8TB iron wolf, and the noise is more than double, and far more annoying in character. Heat and power also way higher. For next upgrade I want 14TB plus disks, and they are all horribly loud. I want high capacity, but low noise, low speed, low energy, since I can have multi TB of NVMe cache.
go SSD....smaller nas with nvme.
RAID 10 with 4 hard drives seems like a good idea.
Why not RAID5? You lose a little performance, but you'll get 50% more capacity.
@@Hotrob_J RAID5 has problems. Do research.
@@brandon_wallace RAID 10 has some advantages over RAID 5/6 including likely rebuild time in case of drive failure for similar/same size drives. I have used RAID 5 for many many years and know that not only are the rebuilds likely to be slow with drive failure but worse is a failed rebuild which is a real issue that does occur. With the price of larger drives going down, a nice 4 drive RAID 10 is very practical.
You missed one point. Kinda. It falls in line with the performance aspect. the larger the drive the longer the rebuild takes on a failed drive. So a 4 or 6TB drive, depending on how full the prior drive was may take 3-8 hours depending on the speed of your system / controller. But a 20TB drive could very well take 24+ hours which in turn leaves your array in a vulnerable state for a second failure. BTW this is why NO ONE should be using RAID5 anymore. The happy medium for me is somewhere between 10TB-16TB drives. They offer larger storage, but with a reasonable rebuild time.
Great information and I love the presentation. You are definitely not boring, even if some would say the subject isn't the height of glamour. I bet you could be a stand up comedian too.
A lot of the useful stuff I have learned about NAS technology is from your videos. Nice one, thanks!
The required amount of redundancy depends on how you plan to use the storage. If the NAS is the primary storage, then redundancy is very important. If the NAS is backing up primary storage on other computers, then RAID might be less critical for some people. The other consideration is point in time, or off-line, copies. If you have a fire, flood, theft or a ransomware attack, it might take out the whole NAS. Therefore, you need protected offline copies. That means maybe double or even more storage. How should we provide that? Does that remove the need for redundancy in the NAS?
Would not want one with Helium because it would eventually leak out then it will fail.
This is absolutely through and incredible information. You just saved me HOURS of research
16tb drives is my sweet spot for storage to drive failure ratio
More drives is always better. If you have one disk drive, failure of that one drive and you could loose everything. With more drives, you can run a raid array. With options for mirrors drives. Options to strip across drives for incredible speed. Or data protection using a drive for bit checking to ensure data stays intact. Just swap out the bad drive. And then the ultimate, use them all together. Speed, reliability. So many options. More is always better.
I've been playing with that RAID calculator while I listened to your analysis, and boy, now I do have a headache.
Here is how I do it: get the price divided by total capacity to get $/TB. That is the true cost of your storage. Then you can compare apple to apple on all of your drive options and pick the cheapest one.
Just note that there is a trade off. The more drives you have, the more power it is going to draw and the more points of failures there are in your system.
Having not watched the video... both have their benefits and drawbacks.
Benefits:
More means higher throughput can be achieved and higher levels of redundancy can be gained making the setup more resilient when it comes to disk failure.
Bigger means less power draw, less vibrations and less potential heat, less physical space used and more capacity.
Drawbacks:
More disks is more power draw, more vibrations more heat production and more physical space used. With the added redundancy comes less capacity as the redundancy means disks are there just to cover the situation where one or potentially more disks fail protecting you from data loss in those cases.
Bigger disks means less options for redundancy as you have less individual disks, less theoretical throughput and often higher cost because even though the cost per GB drops the amount of GB's per disk is significantly higher.
In the end it does not matter much which one you pick as long as you first take some time and think about what your goal is with the setup maximum redundancy and not to concerned about max capacity well more disks is better. Maximum capacity and not to concerned about the data's longevity less big disks is the best option. If your chosen NAS enclosure allows for more disks than you are currently using then less but bigger might also be a good option as it will allow you to grow the storage capacity over time.
Things are even more interesting when looking at CEPH instead of a single NAS. Off course you need at least 3 servers and fast and dedicated network
I enjoyed this video very much! Very informative!!! What I would have like to see is a graph that shows where the flipping point is to decide on more or larger drives, including the NAS itself.
I have an 8 bay DS1821 with 4x 8TB. The rationalisation was that as my data needs grew I could either add more 8TB or on the basis that over time costs would come down start adding say 12TBs and through SHR move to an all 12TB RAID.
So far so good.
I often recommend this very thing for those who have their heart set on a 4 bay but have a little more budget left (or have over specced their day 1 storage needs significantly - slice a TB or 2 off each drive, pump that money into more bays). There are arguments on both sides (i.e drive availability down the line), but if you are running on an SHR, then that's less pertinent.
@@nascompares “87 USBs in a carrier bag…” - I loled!
It takes less time to rebuild the Raid with a smaller HDD. This would be an Argument for more smaller than fewer larger drives.
I just built my own NAS after going through a 2-bay NAS and then added a 4-bay NAS. I built one with 18 HDs and it is much more expandable. I am using UNRAID and it has been great so far. It is much faster and I have so much more capacity. When I need more, I will replace some of my 6 TB drives for 12 TB drives.
I just bought WD Gold 20TB. Crazy fast Built-in NAND and transfer speed up to 500MB/s !!! and much less noise !! I always love larger drive than small one. Because you will save money to throw away small HDD.
The answer is simple: the best is to have lots of big drives!
Crib the storage perspective, of course, not the noise/power consumption.
Of course, with larger drives one should be very sure of the backups. And preferably use 2-disk redundancy to boot. It may be also result in higher ram usage.
Lots of big drives = lots of big $$$
@@nascompares True dat. No other safe way though if you need to keep 1tb of phone pics, several TB of backups and a cr*pload of naughty videos
Actually, I think in most case It depends on how many IO is on your motherboard
such as, if you have 4 SATA ports and you used 2, then you will soon realize if you buy too many low-volume hard disks, you will soon be out of IO to let you upgrade in the future
also, why get an array of Smaller HDDs instead of getting an SSD, SSD is using fewer SATA ports, and still getting the speed
Still not clear which is best 🤔🤔, so I bought a QNAP TS-1232PXU-RP-4G 12bay ($1475) and put in 12 x 20TB drives (Ironwolf Pro $329 each). After RAID 6 have 185TB available 😉😉2 x 10gbe, 2 x 2.5gbe network connection. Bought 13 drives to have a spare. Total cost before tax: $5752.
You were two weeks too late with this video, I just brought 5 16tb drives. I refuse to do the maths to see whether I made the right choice!
Nice Dreamcast sticker.
Cheers mate. It's seen better days though. New PC arrived and debating making it a 'Genesis', 'Mega Drive', 'Master System ' or a 'Saturn'
Wd red are 5900 rpm until 6TB. They are much quieter. The 7200 rpm and above are noisier.
OK Typed up a long post and then lost it all... So here is a short version.
Check the manufacturers HCL (hardware compatibility list) before buying drives. And if you don't find a drive of the capacity you want then consider if it's worth the risk. RAID controllers can be real finicky about drives.
You may feel SATA and SAS is mature tech and there should be no compatibility problems, but there are and there will be more. I've worked with (from memory) Adaptec, Areca, Raidcore, 3Ware and LSI. Sometimes the compatibility problems are blatantly obvious, but sometimes they are a creeping problem that takes time to develop, and they don't get better with time. Sometimes a firmware upgrade of the drives or the controller can help, but there's no guarantee that either is coming if you start out with incompatible hardware.
Also stress test the arrays before your start using them. Run every storage test you can think of on them, and then try some more. Check the RAID logs and take note of any warnings. You don't want warnings! Not even the non critical kind. Make sure there's as little vibrations as possible. Vibrations can play havoc with RAID arrays even if they are not strong enough to cause a head crash.
Also don't use Shingled magnetic hard drives. They are a pain when used for RAID.
Temperature! A interesting paper published by a storage company probably a decade ago showed that the ideal running temp for HDD's seems to be between 35 and 45 °C. Higher or lower temperatures showed increased failure rate. But don't take this as gospel. However we do know that high temperatures are bad in general, and 40°C is a quite easy target for HDD's.
I feel like power consumption… isn’t a factor. You’re going to consume more power. You should expect that.
Also noise… isn’t a factor. HDDs make noise. If you don’t want noise, don’t get HDDs.
Here are my take-aways:
1. Don’t just get 2 drives. Because you’ll end up using one for parity only and waste the space you could use.
2. Get 4-5 smaller drives at once so you can benefit from the performance boost. This also ensures you can have more useable space over all.
Unless you have the cash to fill out 4-5 18TB drives in your NAS, just get smaller drives. Then you can have better performance and more redundancy.
If you didn’t want high power consumption and lots of noise, you shouldn’t be buying a NAS and filling it full of HDDs.
Bigger drives are better. But if the data is important to you, the cheapest way (if it is not for professional needs), get the biggest drive vs price you need, and have an extra one as backup you dont use except for backup. Keep that backup away from power in some storage shelves or so. Hdd you dont use last very long. Had a drive from 10 years ago that I almost never used and put it in 'cold' storage, so unplugged in a shelves, and worked like a Sharm.
Also had a drive that was over 15 years old. Worked aswell
I've been toying with getting an expansion unit for my 1019+ to get more drives or upgrading the current ones I have. 2 - 8 TB, 3 - 4 TB, 2 - 500 GB cache.
Thank you for the thorough video! Perhaps one other aspect to consider is the best way to increase space with a 4-bay NAS that has expansion capability. I have a 918+ with 4x4TB. Shall I replace the drives with 4x8TB drives, or shall I daisy-chain a second 4-bay NAS with another array of 4x4TB?
I think I’d go for replacing the 4x4 TB with 4x8 TB because, well otherwise you need to buy an expansion unit and additional drives. You then can do that again. But at last you have to replace all 9 drives.
*what he said* - Yeah, buy the new larger drives (8TBs are regularly on offer too). Plus, if you can stretch to it, buy a really low end 4 Bay and then set it up as a remote backup device. A proper value one with a Realtek CPU. Then have it on the network or remotely at a friend/colleague/family member's home. Sync them... boom... BACKUP CITY!
@@nascompares Thanks! that was exactly my next step!
@@nascompares I was looking at the DS418 for that purpose
I'm no Synology expert and these types of questions can be best placed upon NASCompares forum.
For what it is worth; we own a lot of NAS and never ever expand a NAS, we either replace the NAS (or HDD's) or place another NAS into the network.
The problem with expanding your NAS is that there is often a catch; either you lose (some?) redundancy, or lose speed & performance.
Plus the risks when something happens between either of them, how will it cope with the problem, will it not destroy/crash the volume?
And the tipping point is often more the costs; what is the budget/how much are your plans going to cost.
Expansion-units are quite expensive, and with the (technical) drawbacks, often the better choice is looking into alternative paths to increasing the capacity..
One idea to consider is the difference in power for say 4 big drives (say 14TB each with 1 parity) versus 12 small drives (say 4TB each with 2 parity).
The comparison of a small size storage ( smaller than 20 TB ) I think was the focus of the video and not larger size storage (35+ TB storage sizes.
Overall the video hit the points I think are important.
you start with saying smaller drives are cheaper, and while they are cheaper as singles, if I were to buy a skyhawk 4tb its 21.5 per tb, a 20tb is 17.2 per tb.
an exos is bigger disparity, in favor of larger drives.
Also remember: an active raid is not a permanent backup solution, it's a stop-gap. You should always do regular backups to an offline media as well. I suggest a raid 5/6 for active use and backups then a mirrored external for offline.
And when backing up remember this law in computer backups “one=none, two=one” the more backups the better. Always good to have offsite backup, onsite backup, and a locked in a bank vault backup.
@@Technichian462 exactly. I try to put my important documents on as many drives as I can. I think my most important documents are backed up over 5 times lol. I even use my old phones as storage. You never know when it'll save you.
@@Technichian462 one question we were asked in our information technology class was "now what do you do if your tax returns are backed up three times in a house that burns down and you can't get access to either of them?"
@@davel4030 Thts why backups are always done/kept off site.
You need to weigh the importance of your data. Critical mission data, data that you cant afford to lose, gets back up every day, maybe even every hour, and it gets backed up sequentially. Indeed, that means you can have hundreds of backups. You label them by date and time. That back up becomes a time machine for your data. Allows you to wipe the system clean, and reload it he system from last week, yesterday, even last months data. If you are ever infected with malware you will wish you had that sort of back up. I’ve seen it happen. Lots of important data that could never be replaced lost forever. Millions of dollars worth of data just gone. Never to be replaced. The company went bankrupt. I dont think the owner ever did anything else after.
All he needed to do was spend a few thousand dollars, to have multiple raid arrays running and offsite back up of that. With backup running every night, with sequential saves going every days for two weeks.
He could have been like Elon Musk, but not like a shyster p.o.s.
Fantastic comparison!
Cheers bud. Appreciated
Appreciate the mention that if you buy a number of the same drive from the same vendor you are likely to get drives from the same lot which if that lot has a problem means your risk for trouble is increased.
A MASSIVELY overlooked (if slightly rare) factor, yeah. Glad you spotted that! Cheers for watching bud
I use 8 x 8tb as standard in my 8 bay nas's, JBOD. Reason is, I use a 2nd and 3rd nas as backups, and if a drive fails I just copy the data onto its replacement, that way I just keep 1 nas running, otherwise 2 would be on all the time, I do a backup using goodsync once or twice a month, I tried synolgys drive sync, too automatic. I like the control of goodsync.
I would say more drives, more chances at redundancy and rpms somewhat combine with raid so the speed increases.
Thats not to say you should go small either, go 8-16tb per drive.
Honestly I just have a bunch of 4tb red drives (every seagate I have ever bought failed within a year.. all of them..) My reasoning is its easier to replace a 4tb drive than an 18tb drive. Both in cost and in time. If I fail on a rebuild then I only loose 4tb of data. (I use unraid) I only backup what I can't get back (pictures, home movies, etc. I can always re rip my dvds and such. 3-2-1 can get expensive otherwise. Especially with larger drives.
The little nas boxes seem pretty neat but frankly an old pc with an hba card is all you need. Buy unraid once (or use truenas, openmediavault, linux, whatever you prefer) works. I prefer unraid because of the way it works. Even if you fail on a rebuild you only loose whats on the failed drive. With raid you loose the whole pool. With nas boxes your upgrade path is kinda expensive. With other options you can just use your old pc when you upgrade.
My 2 cents worth. A lot of options. Depends on risk, time, and finances. Everyone's mileage will vary.
Having just replaced or added HDD's to both my 4 bay Synology and my 4 bay QNAP, my next NAS purchase to replace one of these will be an 8 bay. And I'll install 3 larger HDD's in RAID 5 or Synology Hybrid RAID and add new and larger drives as the need arises and prices fall. The worst thing about smaller NAS is the cost of pulling out perfect good, decent sized HDD's because you need to replace them with larger drives. This is especially problematic with QNAP as you need to have all drives be the same size. With Synology's SHR, drive sizes can be mixed ( although even that is not perfect). The extra cost of an 8 bay vs a 4 bay is easily saved by the ability to continue to use my older, smaller HDD's. Don't be deterred by up front costs; look at longer term use and future purchases and redundancies.
Cheaper more drives but what about power consumption? More watts consumed or is the same? Let’s say will last 6 years and had to pay more electricity ⚡️ during those 6 years that also impact
The "not all eggs in one basket" is a bad figure for RAID5. Speaking from probability of losing your data, using LESS drives is better. Let's say the drive failure rate is 3% per year.
- 2x6 TB RAID1: probability of losing ALL data (2 drives fail same year) = 0.03 * 0.03 = 0.09% per year
- 4x2 TB RAID5: probability of losing ALL data (2 drives failure plus 3 drives failure plus 4 drives) = 0.518 % per year
so the RAID1 is A LOT safer BECAUSE it's using a double safe basket instead of multiple baskets that are connected and ALL fail if 2+ fail
Or, Dell T630 LFF server £350.
Used 2GB cache RAID controller £110
6x used 8TB HC520 drives, £70 each.
Needed a server as well as a lot of storage, so leaving out CPU & SSD upgrades.
I'm just in the processes of putting together a NAS...SOOOOOO many questions and this was one of them 😅
Per gig is cheap now-a-days. However, i don't think i should admit i still use Barracuda desktop drives in my NAS...
Their cheap, compared to RED drives and IronWolf.. Besides,from past experiences, they 'whine' allot in idle mode... Could of just been bad drive, but i doubt. These were 4TB drives
Also, power-saving can make up the difference between buying big drives... The presumption your making is NAS's are designed to be on all the time and active all the time, which is not always true. There is always going sections of 'idle', time, (particularly after midnight),. If you have Scheduled backups going on a QNAP, your gonna allow a few hour either way before the next starts to prevent possible increased failure. In that time space, the dives will spin down after 30 mins (usually) thus saving power. If you work that over a given year, that's still a bit of energy saved right there.
huge enterprise drives can go 300mb/s sequential while a 4 or 6tb drive usually cant even hit 200mb/s especially if they're 5400rpm so fewer disks can be as fast depending on the size difference. Also, the power use is substantial when using more disks. disk power usage can be more than the rest of the entire system combined when talking about 10+ disks
I've got this idea recently.. In theory, what about using classic non-NAS type 2,5" SATA SSD's? For example four 2TB SSD's in RAID 5? I know, it's more expensive. But it's silent, less power consumption and if I would have money for it? Just for home using, not really too much transfering data and important data will be backed up to external drive too. Is it there some statistics or something? What's yours opinions guys? I have few data SSD's in pc for long long time using by daily basis and they are immortal.
This is what i am pondering aswell, fast archive can be well maintend , duplicates removal etc. Some external snapshots for less intresting data etc. Question is for the installed OS drive mirrors, put that to an expensive storage? How many snapahots to keep etc.
Nice video!
I always calculate the price per TB. ;-)
As that way there is often a HD-size (and above) that the price per TB is higher than the smaller/previous-size one. (price per TB)
We often opt for a wee bit smaller but more drives. As we always have 2x cold standby HD's (thus unused!) near the NAS.
But more drives also means more power-use as you already indicated and also more heat. (more cooling might be needed)
Business-wide you should replace your HD's every 3 years (5 years, max) but for the average user at home, you replace when really needed. (or after 10 years?)
To reduce risks, we also opt for multiple NAS. As a NAS may fail (power-supply, firmware-upgrade, bricking)
With larger drives the rebuilding of the RAID also may take longer., when something does go wrong.
BTW, also worth noting, the weight of the NAS might become significantly higher when you are using more drives (noticeable after 8x drives IMHO)
Generally speaking, the 6TB and lower are robusts as rocks (longevity), 8TB to 10TB are often the sweet-spot for pricing in my experiences.
Word of advice: buy as many drives as can fit in your NAS as down the line by the time you want to buy additional drives, the manufacturer may have moved-on to newer models..
There are indeed so many arguments and I do not want to use the words pro's and con's but rather benefits and caveats.
Your video is quite clear there is no definite answer. Plus, I would say, personal preferences.
But excellent to line-out some of things one can think about when making such choices.
As a famous Dutch soccer player once said: every disadvantage has its advantages ...
As always, wonderful observations man. This video was largely born of an issue I faced with a user in the advice section that was buying a 24 bay and wanted to partially populate. I did highlight (as you did) that all too often smaller or more niche drives can go 'end of life' or changed (ANYONE WHO BOUGHT FIRST GEN WD GOLD HDDS - YOU KNOW!), but it wasn't especially heeded. 15 mins into composing my email to the chap and I realized it was effectively an entire video script. The same is true of a UPS video that should go out in the next week or so (though that is even more niche and the pros and cons even more ambiguous). Also, I know this vid is only like 2 hours old, but I'm surprised my points about HDD noise are not getting highlighted more. Sometimes I genuinely feel like I am the only sod who is talking about how bloody noisy ent/pro drives are! Too, TOO many phone/video editors are buying 4 bays with 10GbE, stuffing it with 20TBs and then getting a shock that it sounds like R2D2 having a panic attack! I'm quite looking forward to seeing your thoughts on my video coming soon (genuinely called) '10 Dumb Questions about NAS that are not actually dumb' (part 1 in possibly a 3 part series). Have a great weekend bud.
@@nascompares You most certainly have a very valid point about the noise, that's also why I have not all of our NASses turned-on all the time.
(and some experiences with WOL & 10GBe haha)
And even then, when I do a full shutdown (once so often), the true silence can be deafening!
Although modern-day high throughput switch's fans can, especially when coming of age, become a much larger nuisance in my experience.
That's also why I have the truly sound-dampened 19 inch racks (Ucoustic Edge) in use.
I always look forward to your videos, as they make me think, consider and sometimes even reconsider.
And think outside the (padded) box, being more creative, flexible, and sometimes be amazed. (in the positive sense, of course!)
To me personally the dumbest question is the question not being asked, if you pardon the pun.
Thank you again for this gem of a video!
You too have a great weekend!
It depends. If your nas it is just for fileserver and a few clients go foe bigger disk. If youf nas contains db's and lot of clients, better more disk..
I would have divided the presentation between Consumer needs/suggestions and business needs/suggestions:
in case of business you need IOPs and you do not care about noise because you have a datacenter (for small that it can be), so smaller disks in higher number give you better performance anche noise means Enterprse disks which have higher number of TB/year. at the same time as enterprise, for small that you can be, you should implement the 3-2-1 backup rules or at least having a backup site. the additional point is having a redundant power supply on your NAS.
for the remaining part, what you said is perfect for a consumer site.
The issue/barrier I will always have with tackling any subject with separate Consumer/business POVs, is that it immediately overcomplicates the subject. That's why I make separate videos for home/business. I recommend you watch my Pro Vs Ent drive series
great video, thanks.
Thanks mate for the great info. I have a 5 yr. old TS-653A with 5 WD60EFRX and I'm stuck with 1gbps connectivity. I'd like to get at least 2.5gbps upgrade so I'm going to get another NAS with faster networking and use it as a staging NAS. I'll add a 6th drive to my 653A and move it to my in-laws house and work out syncing between the 2 NAS.
It depends on what you need... More I/O - IOPS vs. need for more space. You can't really get both... Great video! I have done a lot of videos on this very topic...
You talk about the possible more possible points of failure.. but you miss the big point with Raid 5 vs Raid 1... raid 1, if one drive dies.. you take it out.. order a new one.. re-raid it when you get it.. with raid 5.. when one fails.. you have to get another drive and rebuild it.. b ut while that is going on.. the entire raid is OFFLINE.. so I normally recommend.. if you RAID 5.. order a spare to minimize the downtime...
I have HDD's of various sizes. 2TB 4TB, 6TB, and 16TB.
What about the best of both worlds: lots of big drives?
If you have two large capacity drives and one of them fails…that’s a HUGE SLOG to replace in one hit. If you have lots of small drives and one fails it’s not going to hit you as hard when you suddenly have to get the replacement.
Aged like milk storage is getting cheaper and ssds to so soon you can have fast longer lasting drives.
Based on a 6 or 8 bay Synology system, whats the best size drive for reliability in Seagate drives..?
Why not "more bigger hard drives?" When I bought my NAS I got four of the biggest HDD I could at the time. Now they make drives that are almost twice as large, so I'm jonesing for an unneeded upgrade.
The noise is a legitimate concern, but I think for most people the best solution is to try and have a wiring closet / server room type setup and put it in there where you won't really hear it. Of course, the problem with that is that the house basically needs wired for ethernet with a patch panel. I wish I had that.
Again, 'more' 'big drives' is an option, but that's not the question in the video. That's why it isn't really touched on.
This isn’t strictly relevant to this specific video but I’m asking advice I’ve been given the task of assembling a moderately large NAS for a small company.
I’ve decided I am going to include cache but it’s the type to get I’m confused by.
It happens to be a Synology NAS I’ve gone for, and I noticed that specific types of M.2 Sara or nvme are recommended. It basically narrowed it down to WD Red, FireCuda 520 and Synolgy’s own 400 or 800Gb Nvme.
My initial reflex was that it was probably a good idea to go for Synology as it’s the same make as the enclosure but 400Gb of Synology SNV3410 Cache is about twice as expensive as 1Tb of WD Red nvme.
Why is this and is there anything that justifies this price difference?
I prefer drives under 4TB as I find them to be more bulletproof. Bigger video and game files as well as AI models have caused me to now need large capacity drives. I been on a computer since DOS. My first PC had a 100MB IDE drive. I used that drive till IDE was fazed out and SATA became the standard. It never died nor slowed down. But it did sound like the Predator from the movie. I stopped using it cause IDE was just too slow. I am 40 years old and I have drives that I had when I was a teenager that still work fine today. The one that runs my pfSense is a 2.5 inch that has been in a laptop for about 6 years till it went in my Desktop for extra storage for about another 5 years and now runs my firewall. Its the only old drive I have that clicks. Been clicking for years now but will not die. But every singe 4TB or larger drive I get will need to be replaced at some point cause they are sensitive like lil girls.
Vibration or noise or impact or temps or looking at it too long will break it. I have spare large drives just in my closet. No small drives cause they just won't stop. That 2.5 120GB has been in bumpy cars, dropped 100's of times, bumped into, ran sitting upright, ran upside down, sideways, slanted, its older than some peoples children, and still clicks along. Its seen soo many video drivers, windows updates and PornHub. But 4TB and up...a loud noise might startle the thing and make it slap its forehead with the back of its hand as it faints. You gotta wake it back up in the controller. Shaking a grown man and he will most likely survive. Don't do that with a baby. But HDD's is different. You can shake the baby HDD's but if you shake the big grown HDD's they are dead dead dead. We are at a age now where files are big now so I am building out a server rack using large HDD's not because I need a server rack but because the server rack needs a safe place to even be a server or NAS like Hollywood. PC's are built different. They will save mkv's and load steam in the streets of Brooklyn even after being dropped violently because a bee flew in your face and swatting at it and missing the bee made you smack the PC on to the ground. Them small capacity drives are built like 50 Cent
Get more drives, put it in your utility room, connected to your core switch, on a UPS.
It would be really interesting to me if you could make videos at regular intervals for things like When will the price of hard drive come down and when will we see bigger capacities in the small classes.
That's a good idea, but would be quite a lot of videos with quite short shelf lives. That said, I really, REALLY recommend the Price per TB tool that Eddie built (nascompares.com/2022/12/28/how-to-choose-the-best-value-hard-drive-and-best-price-per-tb-get-it-right-first-time/) This will allow you to enter the parameters of drives, capacity, brand, RAID and more - then it spits out drives (priced as of 'now') on amazon. Its updated daily and if you are looking for the best way to track how much TBs your money can buy you in 2023, its the best there is tbh. Hope this helps!
currently my Media collection is currently under 3TB so I do not need a very big set up yet. Currently using 3 2TB reds in a raid 5. About to upgrade to 3 3TB drives in a raid 5 Then have a single 10TB HDD as a Back up.
I dont think we should only look at size vs number but also failure rate. I would rather go with a disk that doesn't fail on me that often then the one i need to buy a new disk every few years and rebuild the RAID.
As of this month(August 2022) it's actually cheaper per gigabyte to buy high capacity drives.
If your Data is very very very important than use a 4 bay or larger, with larger Harddrives in Raid 1. in a 4 bay for example you have 3 Discs of failure.
If you have seriously important data don't use a NAS use a SAN, lots of drives only generally gives a performance increase only for reading not writing, RAID on NAS devices usually has some restrictions based on implementation of the standards of the supplier vs on board raid provided by server manufacturers. NAS providers are great at vendor lock-in. Also make sure you buy drives compatible with the NAS as they don't cover warranty issues otherwise. Generally pro series drives offer 5year warranty non pro are 2-3 years
Great video!
i have 2 slow ones. Is there any way to add faster ones but don't make the whole system as slow as the 2 slow ones?