I just bought a RN516. So now I'm actually in the club of people who spent money on the expensive devices that Netgear sells under the readynas system. I don't want to have to install through SSH a fan system. The beta 3 still has fan issues for legacy units, but I don't care about that. I want to know when specifically I can choose what fan speed the fans will run at. I don't care how quiet the systems are, I want to be able to manutally control how fast the system spins the fans.
I know other people have requested it numerous times... I'd like an official comment on when we can expect to control fan speeds. I don't want drives getting too hot because of those that whine about how loud the unit is. I want the option to be able to spin the fans as fast as I see fit to achieve the desired drive temps that I deem acceptable.
Is this something we can expect to see in the 6.4.x platform or are we going to be waiting much longer than that? Do we have to wait until new units are produced that aren't even running 6.x software before we can expect to see it? If it's too much to integrate why can't an app be made?
People want to know and as an owner of an expensive 6.x unit I want to know.
Best of luck with this. My impression is that Netgear only worries about features for "the masses" and doesn't care about advaced features for advanced users. Like Microsoft, they now think they are smarter than me and know what I should want instead of allowing me to make the decisions. Better control over alerts is yet another example.
Well my box locked up hard over the weekend (probably due to temps) and now it won't boot back up again (gets stuck on booting.) Since it locked up hard I imagine the drives went over the failsafe temp which would have caused the unit to shut off, but it was locked up and thus that didn't happen. Not to mention that the fan wouldn't spin up any faster because again, the device hard locked up.
I've got 6 8TB drives... yes they're all under warranty but this just goes to show you that whatever fan control system they have in place isn't working and now I'm probably going to be out significant money, or at least $60 sending the drives back to seagate and praying they warranty them out for me without any guff.
Absolutely stupid way of doing things. The machine wasn't even doing anything intensive. I was doing one ftp transfer that was running at 30mbps and a time machine backup over wan so that was going at 4mbps. That's all it takes to hard lock up a readynas box.
I think at this point I should just sell all 4 units I have and go elsewhere. They clearly don't have this system dialed in anywhere it needs to be and it sickens me to think of the people that spent so much money on these devices only to have them be so unreliable and unstable. How is it possible?
Don't even ask. They aren't on the HCL so it's obviously my fault that the box locked up.
Just FYI I decided to wipe out the unit and have it factory default and all the drives are perfectly fine. I had 95% of the data backed up (was in the process of backing up the other 5% to an offsite backup but that's when it locked up.)
I'm just going to find a different 6 or 12 bay unit from another manufacturer and sell what I've got. I've tried to live with the bugginess for long enough.
You've made your opinion on drive selection very clear. However, if you would have done a little research before purchasing them, you would have surely run across a Linux kernel bug report (a report that now has 80 comments) where many, many people are seeing all sorts of bad interactions from these very disks -- on all sorts of different hardware platforms, and across all Linux distributions. These drives are simply well-known to be problematic on Linux currently.
Regarding fan control, I posted some info just over a month ago here, which explains how to set an arbitrary minimum fan RPM, and also confirms that adjustable fan control is indeed coming. It will definitely be added to 6.x firmware.
You've made your opinion on drive selection very clear. However, if you would have done a little research before purchasing them, you would have surely run across a Linux kernel bug report (a report that now has 80 comments) where many, many people are seeing all sorts of bad interactions from these very disks -- on all sorts of different hardware platforms, and across all Linux distributions. These drives are simply well-known to be problematic on Linux currently.
Regarding fan control, I posted some info just over a month ago here, which explains how to set an arbitrary minimum fan RPM, and also confirms that adjustable fan control is indeed coming. It will definitely be added to 6.x firmware.
To address your 1st point... there are thousands of these drives out in the wild. The only correlation I can make is if I am to agree with your sentiment on the drives not behaving with linux based on 80 comments, that based on the comments on these forums about 6.4.0 it looks like your firmware was incompatible with many of our nas boxes. (And I'm sure at this point they've sold more of these drives than you have of the nas boxes, not sure but I'd venture a small wager that that's the case.) There have to be at least 80 comments and different threads on this forum complaining about 6.4.0 and the fact that it's still be offered as the latest non-beta is crazy. Look at how many people were willing to back up their device and downgrade to 6.2.4 or lose their data entirely just to get a stable unit. So don't tell me jack about stability about these disks when stability regarding your firmware isn't any better!
Getting to your minium arbitrary fan RPM setting, it doesn't show up on my nas. SSH'ing into my unit there is no min_fan_speed_override. And yes there you state that in the future there should be an app or native solution. I read it, I just don't believe it because while doing other research based on 10gb ethernet being available on the readynas rn516, I found a thread where the specsheet said that netgear was going to be releasing an add-on, and a year later no add-on has been provided and no response from any official netgear employee has come about face. https://community.netgear.com/t5/New-to-ReadyNAS/10G-on-RN516/td-p/905587
Please... insult me some more! I had no issues with 6.2.4 but I keep hoping that every new beta that comes out only fixes things and doesn't break 10 other things in the process but when my RN516 shows up tomorrow I think I'm just going to downgrade it to 6.2.4 (unless that's what's on it) and start over from there and stay there until I hear your firmware is actually compatible with the box. I had zero lockups, drive issues, anything to complain about on 6.2.4. I tried 6.4.0 betas and then made the utter mistake of upgrading to 6.4.0 when it was official. In my laziness of not wanting to downgrade I stuck with it until the betas came out. While they showed improvement to get us back on the right track, they still aren't as stable as 6.2.4 was and then they happen to ruin my raid array. But hey, it's the drives, they did it. That's why for like 4 months they've had no issues with read or write in raid aside from lockups which only occurred when 6.4.0 came out.
The 8TB disks are using new technology. Support for disks in Linux relies on development which can lag behind support in e.g. Windows.
Furthermore archive disks are not intended by the manufacturer to be used in RAID arrays. Indeed a review (http://www.storagereview.com/seagate_archive_hdd_review_8tb) tested these drives in a different NAS and found the RAID performance of these drives to be terrible.
Poor performance with these drives is not new. Now it is possible you may have run into one of the common issues with 6.4.0 which may well have been exacerbated by your drive choice.
We are working hard on a 6.4.1 update to address common issues which some users have come across and have a beta avaialble for those who would like to try it. Feedback here and via support is very valuable.
Inevitably with a firmware update some users will encounter issues. This is unavoidable, particularly with a major update. 6.4.0 is the biggest update since the very first OS6 release. When a major update e.g. 6.4.0 is released (particularly one which you cannot downgrade from) some users will choose to hold off updating for a firmware release or two.
I am running 6.4.0 on my 516 with 6x6TB WD REDs and my unit is running fine.
There will be no 10gig card for the RN516. There is 10gig on the RN716X.
Override files do not exist by default (that's why they are called override files), but you can create them e.g.
I know we have differing opinions over 8TB SMR drives. I posted the link to the linux bug report initially - it probably is a fair comment that we should have looked first BUT I think its an ever fairer comment that the Netgear devs should have twigged that when upgrading the kernel in 6.4.x and I do think, although millions may disagree, that Netgear should buy an SMR drive and test it - as I did offer for a while and didn't hear anything - and unfortunately had to use my third 8TB drive in my JBOD backup array (which is rock solid on 6.2.4). I know there are potential issues in RAID arrays but its also cheap storage and *can* be stable (as demonstrated in 6.2.4). Unfortunately I'd happily test if I didn't have to lose everything going back to 6.2.4 and I can assume from ifixdevices post that 6.4.0 still isn't working right for SMR drives.
I am pleased that all the other issues i've had with 6.4.x (network copy timing out, 104 hanging on shutdown etc..) have all been reproduced and fixed, and still happily own 6 readynas's (although if anyone can solve my readyCLOUD on Pro Pioneer issue i'd be very grateful!).
Netgear devs should have twigged that when upgrading the kernel in 6.4.x,,,
I don't disagree, though I think the kernel upgrade was needed for other reasons. Of course they have never claimed support for the drives, so it likely would have been rough sledding to propose slowing down new kernel/btrfs features. But it would have been easier for you if you had known up front that 6.4.x wouldn't work with the drives.
On kernel upgrades generally, they take some bullets no matter what they do. If they don't stay current, people start complaining that the kernel is outdated. On the other hand, when they do take new stuff, there predictably are some issues, esp. with third-party apps in the mix. Obviously 6.4.0 turned out to have issues that caught them by surprise. But on the whole, I think they do need to keep up with new kernels.
I know we have differing opinions over 8TB SMR drives.
I think they have their place. Seagate has invested some money in creating better linux support for these drives (https://github.com/Seagate/SMR_FS-EXT4) but unfortunately that isn't nearly finished yet (and a lot of it seems to be ext specific). Note RAID support is on the Seagate roadmap, but still in the future.
25% cheaper per TB does make them tempting. Though I think SMR is likely to end up as stopgap - once HAMR drives show up, interest in SMR will likely disappear. So there's a fairly short market window (2-3 more years), and limited time for Seagate to get the OS support solid. By comparison, PMR will likely have a 12-13 year run.
The 8TB disks are using new technology. Support for disks in Linux relies on development which can lag behind support in e.g. Windows.
I understand they are using new technology internally and like WD drives in the past certain things need to be tweaked for better performance and support.
Furthermore archive disks are not intended by the manufacturer to be used in RAID arrays. Indeed a review (http://www.storagereview.com/seagate_archive_hdd_review_8tb) tested these drives in a different NAS and found the RAID performance of these drives to be terrible.
Has netgear actually tested them in an array? I had a 6 drive array. Could still write large files at 80MB to 100MBps and read large files just as fast. Smaller files undoubtedly gave slower performance, but that's what to expect with a budget drive. You can soup up a Honda but it still isn't a bugatti veyron and neither is in the same price range so why would you expect the same performance. I need to store large amounts of data with redundancy and performance isn't a priority for me.
Furthermore, obviously seagate doesn't recommend them in Raid. They still sell enterprise drives. They want people buying the enterprise drives... these 8TB drives are just an excuse for people who don't need 8TB's to buy a drive so they can further tweak the technology and improve on it (I.E. we're guinea pigs and they get to see the failure rate RMA drives and learn from them.)
Poor performance with these drives is not new. Now it is possible you may have run into one of the common issues with 6.4.0 which may well have been exacerbated by your drive choice.
Again poor performance is not what I'd call running these drives. You have to be more patient, but it's nothing earth shattering. Again I think the loudness of those who think they can spend $200+ on an 8TB drive and expect it to perform like it's $800+ counterpart whine the most and are most vocal about the issue.
We are working hard on a 6.4.1 update to address common issues which some users have come across and have a beta avaialble for those who would like to try it. Feedback here and via support is very valuable.
I have been testing out the 6.4.1 betas. I was on beta 3. I believe my issue was caused because I did a scrub, then balance, then defrag (never saw anything in the notification about a defrag, wasn't in the log.) The first time it had been restarted was after not seeing anything about the defrag so I believe something got screwed up in that regard. It wasn't the drives fault. I have no errors with the drives either via smart.
Inevitably with a firmware update some users will encounter issues. This is unavoidable, particularly with a major update. 6.4.0 is the biggest update since the very first OS6 release. When a major update e.g. 6.4.0 is released (particularly one which you cannot downgrade from) some users will choose to hold off updating for a firmware release or two.
Right but most people don't know that. I upgrade everything to the latest and greatest. I was content with 6.2.4 because it was stable unlike the 6.4.0 betas I had tried but when the OS6 6.4.0 update came out, I thought it was at least more polished than what it turned out to be. Regardless those who could fix it and downgrade did and are waiting for the stability improvements. The key thing in this I can say though is these are EXPENSIVE units that netgear sells. In an enterprise setting upgrading to 6.4.0 might have caused some network admins to have heart attacks (any good network admin would internally test out 6.4.0 before releasing it into the wild on a production setting, but what about the ones who didn't have two units to choose from (again because they're so expensive.)
I am running 6.4.0 on my 516 with 6x6TB WD REDs and my unit is running fine.
Great. What do you specifically do with it and what connects to it? Do you use it with Mac computers? Do you use time machine, ftp and transfer things back and forth between different boxes on different network providers (without using readynas replicate.) I use mine also for running parallels from and I have no issues (it's not even slow.) But I'm curious what you do with your unit vs. mine to see why yours performs flawlessly and mine doesn't based on the lightest of tasks.
There will be no 10gig card for the RN516. There is 10gig on the RN716X.
Why does it have an expansion slot on the motherboard? I mean basically the 716x just has a faster processor (one I plan to upgrade to as my 516 that comes in today won't have any warranty on it) and 16GB's of memory vs 4 and the 10gig card. I think someone in that very post found an aftermarket solution to make it work. I don't know why netgear couldn't sell an upgrade kit but I'm assuming it has to do with the favorite term thrown around "not user serviceable." I'm betting if you removed the riser card and 10gig card from the 716 and put it in the 516 it'd work fine.
Override files do not exist by default (that's why they are called override files), but you can create them e.g.
Did not know that. Something I can admit I don't know everything. It wasn't clear from the directions that that's what you had to do (which is why I'd love an app or native fan control support.)
If you like you could send in your logs (see the Sending Logs link in my sig).
Already wiped out the unit. Figured it was easier to start over as most data was backed up.
Yes, though there are several other threads on various distro forums. Of course we'd like them to work as well as possible, so we've been incorporating some of the proposed fixes along the way to help users that went ahead and purchased them. But it's not realistic to expect them to ever appear on the ReadyNAS HCL, when the manufacturer clearly says not to use them in a NAS right on the data sheet.
I understand they are using new technology internally and like WD drives in the past certain things need to be tweaked for better performance and support.
They are still fairly new, and the linux community couldn't get started until the drives showed up in the general market.
I'm guessing here, but I suspect Seagate will need to provide a firmware update. From a driver perspective, these disks are supposed to work like PMR, but with different performance and power characteristics. But reading through the kernel bugs, they clearly aren't always doing that.
I doubt its a drive fimrware problem as its fine with 6.2.4 and linux 3.0x kernels.
I believe I'm right in saying that SMR drives can be drive managed or host managed. With the former the drive manages its internal structure itself and should funciton exactly like any other drive, with the latter the OS takes some responsibilty for 'managing' writes to the drive - I may be wrong but I wonder if by definition 6.2.4 falls in the former as it is SMR unware but 6.4.x is trying and failing at the latter - surely the easiest answer for now, assuming this is the case, is just remove any SMR specific code from 6.4.x.
I doubt its a drive fimrware problem as its fine with 6.2.4 and linux 3.0x kernels.
I'm not arguing that it isn't stable with 6.2.4. Clearly you (and ifixidevices) found that it was. And its clear that it isn't stable with 6.4.x.
Since the SMR is drive managed, it's supposed to behave exactly like a normal drive (with the obvious performance and power differences). Just that looking at the kernel bug thread shows several instances where the drive is not behaving properly to normal ATA command sequences. By definition its not supposed to ever do that.
That sounds like there's a drive firmware problem in there somewhere. But we will see.
...I believe I'm right in saying that SMR drives can be drive managed or host managed. With the former the drive manages its internal structure itself and should funciton exactly like any other drive, with the latter the OS takes some responsibilty for 'managing' writes to the drive...
There are three categories actually.
-host-managed. The host takes full responsibility for maintaining the SMR. So if host wants to write to track X, then it ripples the write to the end of the zone itself, in order to avoid destroying track X+1. This drive isn't host-managed.
-drive-managed. The host treats the drive like any other drive. The drive firmware is fully responsible for maintaining the SMR properly. I believe that's we have have with ReadyNAS + this drive.
-drive-managed but host-aware. In this hybrid scenario, the drive is still responsible for maintaining the SMR, but the host knows it's an SMR, and has some optimizations to improve performance. For example it might adjust the file system so that it allocates free space near the end of the zone whenever it can. Or it might re-order writes somehow to minimize the performance impact.
It might not help. Something else in the kernel upgrade could be responsible for the destabilization, and the SMR-specific change (whatever it is) simply might not be helping (but also isn't hurting).
But it seems obvious that its worth a shot. Maybe PM skywalker and ask if that's easily done, and offer to test it. (or maybe ask generally if there's anything you can test that would be useful to them on this).
I'm more than happy to send the logs constantly if I knew it'd make a difference. I don't do much with my boxes in all honesty. Time machine backups and file transfers mostly (mostly for just storage of frequently accessed files.)
one box I use for dlna to serve to my TV's and a backup of the main box but again that's about it. Still been having random lockups, had a drive fall out of the array even though I'm not seeing any regular errors that show up in the main smart data.
6.4.1 Beta 4 will have a backport of this fix, which has been reported t successfully resolve the issue by several Linux users experiencing issues with SMR drives.
That's awesome. I'm not really seeing any performance issues at the moment aside from random lockups (which may be caused by the drives, who knows.) I was toying with going with 6TB greens for a while and I was still having random lockups (albeit faster transfers) but I decided it'd be too much hassle to try to find enough 6TB drives at the price point I want and sell 7 8TB drives without losing any money (if eBay didn't take 10% from you it'd be a different story!)
So because of that I decided to stick with the 8TB drives even though they are slower. As soon as charter gets faster upload (which who knows when that will be) I'm still contemplating buying a 12 bay unit and using it for backups for cloud based storage (IE time machine backups in the cloud for all of my mac customers.) The real value in it is that while I know blackblaze and crashplan and other providers are out there, the amount of time it takes to pull data back from them is where you find their value quickly diminishes. Then there's the old "we'll shove it on a drive for you and mail it out to you for like $200) option that's BS.
My idea was to put the 12 bay unit in a local datacenter, pay for 100mbps unmetered data upscaling as necessary and then put in enterprise 8TB drives (only 2 to begin with until I got enough clients to fill those, then add additional drives as needed.) That'd be crazy expensive though, so I'm doing it right now with an RN516 with 4 8TB archive drives. I've got a business fiber line with 30/30mbps and back up clients macs for $120 a year up front with 250GB storage, $180 for 500GB, $200 for 750GB and $250 for 1TB space (with certain customers getting a ltitle bit of a discount.) If possible I do an initial backup (usually when they have their mac in for service) and then it continues backing up to the server when they go home. Then their time machine sparsebundle gets backed up to another nas unit at my home for redundancy. Certain customers get backed up to another location (my parents home.)
The upside for my customers is that if anything goes wrong with their hard drive or computer they get a discount on the repair and I can have them up and running in a few hours. No limits to how many times you can bring in your computer. Back up as many macs you want to as long as they fit the space they're paying for.
