Seagate’s Second Gen Mach.2 Drives Are as Fast as SATA SSDs

kac77

2[H]4U
Joined
Dec 13, 2008
Messages
3,318
Seagate’s Second Gen Mach.2 Drives Are as Fast as SATA SSDs

"These days most people don’t pay much attention to spinning rust. Everyone is using SSDs now, so you don’t see many headlines about hard drives anymore. However, Seagate has just announced what amounts to the fastest hard drives ever made. They’re so fast, they’re equivalent in throughput to a SATA SSD."
 
If they're price competitive, I'll gladly be using these to upgrade my spinning drives when that becomes necessary. I need that good $ per TB though!
My pricing on the 18TB gen 1 units is under $500 CAD, adjusting for inflation and increased complexity I would expect these new ones to be just north of $600 for the same 18TB (technically 2x9) of storage.
 
Well damn, still cheaper to get a slower 16TB or so and use a SSD Cache for my uses then.
 
Well damn, still cheaper to get a slower 16TB or so and use a SSD Cache for my uses then.
I have their Gen 1 units in my, backup, backup appliance. The only thing I need there is good sequential write speeds and storage space because a LOT needs to fail before I get to that box, they haven't given me a hard time yet, the rest of the box they are in, however ... yeah. But the drives are doing what I need.
 
Equivalent throughput to a SATA SSD, maybe. Access times/latency will still pale in comparison.
True but for bulk storage that isn't such a big deal, particularly if the server itself does heavy caching. There's value to be had in cheap magnetic drives that are fast. Not for desktops really (hence these being SAS) but where you want a big 'ole storage pool that doesn't cost too much.
 
Seagate’s Second Gen Mach.2 Drives Are as Fast as SATA SSDs

"These days most people don’t pay much attention to spinning rust. Everyone is using SSDs now, so you don’t see many headlines about hard drives anymore. However, Seagate has just announced what amounts to the fastest hard drives ever made. They’re so fast, they’re equivalent in throughput to a SATA SSD."

Sequential I believe. Random I don't.

And random is what matters.
 
Sequential I believe. Random I don't.

And random is what matters.
Random matters most for machines that are doing active work, but 99% of active cases these are still far too slow for modern workloads. Seagate sells and advertises the Exos line for high speed media storage. So large single file read and write to multiple users where random doesn’t matter much.
 
Its 2022, Hard Drives should be archival.....they shouldn't be operational, we have better parts for that (see comments above, namely anything solid-state). While I applaud this, I'm not sure who the market is....if all drives eventually become this technology, then cool, but between you, me and the doorknob I've been just fine with 5400 or 7200 rpm spinning platters for a decade now. Sort of a solution for a problem that kinda doesn't exist.
 
meh, basically Raid 0 in a single drive.. double the possibility of failure.

I would only consider purchasing after these have been on the market / in use by others after 3 years and proven to be at least as reliable as current gen drives.
 
Its 2022, Hard Drives should be archival.....they shouldn't be operational, we have better parts for that (see comments above, namely anything solid-state). While I applaud this, I'm not sure who the market is....if all drives eventually become this technology, then cool, but between you, me and the doorknob I've been just fine with 5400 or 7200 rpm spinning platters for a decade now. Sort of a solution for a problem that kinda doesn't exist.
All depends on what level you're at. If you're managing a data server farm then yeah, you can afford 5 digits to 7 digits on servers/drives.
I've been looking at building my own home server for a long time, in which both throughput and size matters. And frankly the cost of getting around 160TB of spinning rust vs 160TB of SSDs is incredibly different. I think there is a lot of market for faster spinning drives for small business applications. And I also think that even if you do only want these for archival, no one is going to complain about a 2x speed increase.
The question there still becomes cost vs performance and whether or not it even makes sense. If Seagate can price these at the same cost as other spinning drives with the same density, then that's a win no matter how you look at it (other than perhaps Seagate's reliability. But I'm sure BackBlaze will let us know how they do in a year or so).
 
Last edited:
I wonder if WD will come up with something similar? I admit a bit of a bias against Seagate given past performance, at least for enthusiast/SOHO (I remember the 3TB phenomena which was bad across the board but even worse for Seagate) etc... so maybe these are exempt. I still have a few WD VelociRaptors kicking around from back in the day (for anyone not old enough, these are 10K RPM SATA drives that, prior to the introduction of SSD were the fastest thing an enthusiast could get 15 or so years ago using SATA connectors. Speaking of SAS... do modern at least high end consumer mobos support SAS drives? I remember back in the day some of the HEDT boards used to support it and since its so close to SATA and have gotten cheaper I wonder if its well supported today in X570 or X670E etc?
 
Its 2022, Hard Drives should be archival.....they shouldn't be operational, we have better parts for that (see comments above, namely anything solid-state). While I applaud this, I'm not sure who the market is....if all drives eventually become this technology, then cool, but between you, me and the doorknob I've been just fine with 5400 or 7200 rpm spinning platters for a decade now. Sort of a solution for a problem that kinda doesn't exist.
Let me know when 16+TB SSD drives are sub $200 so I can start using them in my media server and phase out my 100+ TB of spinning drives.
(Which will likely become backup, because most of them have multiple years of life left in them.)
 
Let me know when 16+TB SSD drives are sub $200 so I can start using them in my media server and phase out my 100+ TB of spinning drives.
(Which will likely become backup, because most of them have multiple years of life left in them.)

A media drive and an archival drive are pretty close.

You don't really need any kind of real performance from a media drive.

Even a beefy 4k ripped blu-ray with the highest bitrate multichannel sound is usually only about 45GB for about 2 hours of content, so we are talking about 7.5MB/s.
 
All depends on what level you're at. If you're managing a data server farm then yeah, you can afford 5 digits to 7 digits on servers/drives.
I've been looking at building my own home server for a long time, in which both throughput and size matters. And frankly the cost of getting around 160TB of spinning rust vs 160TB of SSDs is incredibly different. I think there is a lot of market for faster spinning drives for small business applications. And I also think that even if you do only want these for archival, no one is going to complain about a 2x speed increase.
The question there still becomes cost vs performance and whether or not it even makes sense. If Seagate can price these at the same cost as other spinning drives with the same density, then that's a win no matter how you look at it (other than perhaps Seagate's reliability. But I'm sure BackBlaze will let us know how they do in a year or so).

Let me know when 16+TB SSD drives are sub $200 so I can start using them in my media server and phase out my 100+ TB of spinning drives.
(Which will likely become backup, because most of them have multiple years of life left in them.)
Understood, but pretty sure your media streams just fine off 5400 rpm drives, no?
 
Understood, but pretty sure your media streams just fine off 5400 rpm drives, no?
Depends on the media and how many people are accessing the drive simultaneously.

My highest quality files are around 90mbps.

However, raw 4k60 files (from capturing gameplay, etc.) Is easily 100+ mbps, i record to SSD because I've had issues before.
 
I wonder if WD will come up with something similar? I admit a bit of a bias against Seagate given past performance, at least for enthusiast/SOHO (I remember the 3TB phenomena which was bad across the board but even worse for Seagate) etc... so maybe these are exempt. I still have a few WD VelociRaptors kicking around from back in the day (for anyone not old enough, these are 10K RPM SATA drives that, prior to the introduction of SSD were the fastest thing an enthusiast could get 15 or so years ago using SATA connectors. Speaking of SAS... do modern at least high end consumer mobos support SAS drives? I remember back in the day some of the HEDT boards used to support it and since its so close to SATA and have gotten cheaper I wonder if its well supported today in X570 or X670E etc?
No, consumer systems completely ignore SAS. You'd have to get a HBA to support them. Thankfully those aren't too expensive, unless you want hardware RAID and even then. But I've never seen a consumer system that has onboard support for it. You have to get a workstation/server board. It was less than $200 to get an external 12g HBA when we bought one (for tape) though, and I imagine internal ones are even cheaper.
 
No, consumer systems completely ignore SAS. You'd have to get a HBA to support them. Thankfully those aren't too expensive, unless you want hardware RAID and even then. But I've never seen a consumer system that has onboard support for it. You have to get a workstation/server board. It was less than $200 to get an external 12g HBA when we bought one (for tape) though, and I imagine internal ones are even cheaper.
I have a Seagate Exos X 2U12 in my rack as the oh fuck backup, I need a LOT to go wrong before I get to that box but it can't saturate the 10G connection I have running to it, replacing the X16s I have in there with the 2x18's would easily accomplish that task and would be something I could consider down the line.
 
Last edited:
So, everything about this suggests that these are essentially just two striped low profile hard drives in the same case, sharing the same spindle motor.

Which judging by the fact that a single Exos series drive pushes about 270MB/s read speeds makes sense.
 
Understood, but pretty sure your media streams just fine off 5400 rpm drives, no?
Not sure if you’re aware but you quoted two different people. I need a server fast enough to edit from and potentially support multiple people doing so at once.

I plan to have SSD’s in the pool because although RAID 10 of 12-16 drives is great. Even more speed is always good.
 
Last edited:
Not sure if you’re aware but you quoted two different people. I need a server fast enough to edit from and potentially support multiple people doing so at once.

I plan to have SSD’s in the pool because although RAID 10 of 12 drives is great. Even more speed is always good.
Production needs solid state at this point all the way 100%, but when you are done and you just need to archive that raw footage and store the completed files, spindle is fine. And if you want something to do nightly backups to from that production storage it's even better, especially if that backup is done in the form of a system image and not random files.
 
So, everything about this suggests that these are essentially just two striped low profile hard drives in the same case, sharing the same spindle motor.

Which judging by the fact that a single Exos series drive pushes about 270MB/s read speeds makes sense.
Actuator LBA mapping
Exos 2X18 is a device that contains two actuators which can read and write concurrently. The first 50% of the device's LBA space is mapped to the primary actuator and the second 50% of the LBA space is mapped to the secondary actuator. Exos 2X18 supports the Concurrent Positioning Ranges log page 47h which indicates the sets of contiguous LBAs within each actuator for which positioning is possible at the same time. See ACS-5, Rev. 6 or beyond for details. For your convenience, the Concurrent Positioning Range log page details can be identified using SeaChest Utilities software available at www.seagate.com or openSeaChest via GitHub.com .
 
Production needs solid state at this point all the way 100%, but when you are done and you just need to archive that raw footage and store the completed files, spindle is fine. And if you want something to do nightly backups to from that production storage it's even better, especially if that backup is done in the form of a system image and not random files.
Speaking as someone that “has to” edit off spindles everyday, I disagree. There’s a difference between optimum and “needs to”. 4 drives in raid zero is plenty for most video editing tasks.

Although if you are only editing small projects, yes a USB-C or Thunderbolt NVME drive will give nicer performance if you’re okay with having to archive constantly.
 
Actuator LBA mapping
Exos 2X18 is a device that contains two actuators which can read and write concurrently. The first 50% of the device's LBA space is mapped to the primary actuator and the second 50% of the LBA space is mapped to the secondary actuator. Exos 2X18 supports the Concurrent Positioning Ranges log page 47h which indicates the sets of contiguous LBAs within each actuator for which positioning is possible at the same time. See ACS-5, Rev. 6 or beyond for details. For your convenience, the Concurrent Positioning Range log page details can be identified using SeaChest Utilities software available at www.seagate.com or openSeaChest via GitHub.com .

Yep. Which is a form of paralellism. Much like striping, but in one drive.
 
meh, basically Raid 0 in a single drive.. double the possibility of failure.
You made a great point here, that I think people should see. Actuators are essentially what makes a drive fail and this has two.

Also, I have used Seagate many times over, but there last batch of drives were loud, grindy and ran hot (8TB Ironwolf Pro?). Lots of reports of premature failures. I returned them for WD Red Pros.

Additionally, aren't many of their drives now SMR? That has all kinds of performance disadvantages.

Seagate needed to fully embrace the SSD future, before getting left behind. I am sure the bureaucracy would not let them.
 
Seagate needed to fully embrace the SSD future, before getting left behind. I am sure the bureaucracy would not let them.
Seagate makes SSDs, Nytro is the silly name they gave their data center SSDs. They are just mostly enterprise focused, not as much on the consumer market. This drive likewise is an enterprise drive, hence SAS. There's still plenty of demand for magnetic storage in the enterprise arena because it is much cheaper for large amount of storage. You can very well have lots of data that you don't need quick random access to, but you still want online (ie not on tape) and able to access quickly. That's where you would see drives like this.

Wouldn't surprise me if companies like NetApp start using them for their bulk storage servers.
 
Seagate needed to fully embrace the SSD future, before getting left behind. I am sure the bureaucracy would not let them.

As others have mentioned, Seagate DOES have an SSD line.

They are even part owners of Kioxia, which makes their flash memory for them. SandForce is also a wholly owned subsidiary of Seagate these days, and they used to be the premiere SSD controller about a decade ago. They also have access to controllers through Kioxia.

This puts them in a class above most SSD brands which are just rebranded stuff.

The very top of the SSD market is easily in the hands of Samsung and to a lesser extent Intel's old SSD division which is now owned by SK Hynix.

I'd consider Seagats SSD's in a second tier just below these two, but above most "gamer" "badge engineered" SSD's. Others in this second tier category include WD, Crucial, and Micron, and even to a lesser extent Dell.

Most people seem happy with their 3rd tier "gamer brand rebadge" drives though.


SSD's are certainly the long term future for almost all applications.

Medium term - however - hard drives will still be around for mass storage for some time. Maybe another decade or two.

Thankfully the days of being forced to boot and run your OS off of a hard drive are mostly over though.
 
I still have WD Raptors in raid array in a older gaming computer.

Impressive. How many hours do those bad buys have on them now?

I just got done retiring twelve 10TB Seagate Exos X10 drives that have been running for ~41k hours 24/7.

Trying to decide if I want to try to sell them. Not sure if anyone is interested in high milage drives.
 
Impressive. How many hours do those bad buys have on them now?

I just got done retiring twelve 10TB Seagate Exos X10 drives that have been running for ~41k hours 24/7.

Trying to decide if I want to try to sell them. Not sure if anyone is interested in high milage drives.
If the error rates are low there could be interest.
 
Impressive. How many hours do those bad buys have on them now?

I just got done retiring twelve 10TB Seagate Exos X10 drives that have been running for ~41k hours 24/7.

Trying to decide if I want to try to sell them. Not sure if anyone is interested in high milage drives.

I'd take a couple if cheap
 
Speaking as someone that “has to” edit off spindles everyday, I disagree. There’s a difference between optimum and “needs to”. 4 drives in raid zero is plenty for most video editing tasks.
71gogs.gif


Bonus points: The previous Exos 2x14 is, like, 2.5x the price of the ordinary single-actuator version. I feel like the window where this makes financial and performance sense is vanishingly small...
 
View attachment 528136

Bonus points: The previous Exos 2x14 is, like, 2.5x the price of the ordinary single-actuator version. I feel like the window where this makes financial and performance sense is vanishingly small...

Yeah, at that price, just stripe two of the ordinary versions.

Or stripe a few mirrors if you need redundancy.
 
Impressive. How many hours do those bad buys have on them now?

I just got done retiring twelve 10TB Seagate Exos X10 drives that have been running for ~41k hours 24/7.

Trying to decide if I want to try to sell them. Not sure if anyone is interested in high milage drives.

No Idea my son uses that computer. It’s a 4th gen i7 runs just about 24-7 for the last 5 years.
 
Yeah, at that price, just stripe two of the ordinary versions.
Might make sense if you really want more performance, with lots of storage, but can't add any more disks. Seek times might be somewhat better, because each actuator only covers half the disk, so worst case seek is only halfway across. Chances are there's a lot of commands queued though, which helps the disk find things to read on the way.
 
Might make sense if you really want more performance, with lots of storage, but can't add any more disks. Seek times might be somewhat better, because each actuator only covers half the disk, so worst case seek is only halfway across. Chances are there's a lot of commands queued though, which helps the disk find things to read on the way.

Uh...no. Seek time doesn't change. It's not two actuators scrubbing the same platters; each actuator is dedicated to half the platters.

Seagate-HDD-Tech.gif


So, seek time is the same, but you get double the IOPS and throughput. At the expense, of course, of segmenting your storage in half and risking weird and novel failure modes. Would you trust data to a drive that appears half-crashed and half-functional?

Basically, this drive fills a very narrow niche. And if you think you need to fill that niche... you've probably made some questionable decisions at some point. Wrong storage topology; wrong chassis; wrong controller -- something isn't right.

Better to go solid state, or remember that the "I" in "RAID" is for "Inexpensive". Cramming more failure modes into an already brittle system tends to go poorly.
 
Back
Top