The three major HDD manufacturers are selling slower drives, without telling us

[Kilrah]

Again the specs in the datasheets do not reflect it in any way. No mention of SMR, and no characteristic that allows to know that it is. The only characteristics that change are the amount of cache and the "maximum transfer speed", which are specced higher on the SMR drive - but you have no way of knowing that while read may be a bit faster than on the older drive write will be 2x slower.

[mr moose]

9 minutes ago, Kilrah said:

Again the specs in the datasheets do not reflect it in any way. No mention of SMR, and no characteristic that allows to know that it is. The only characteristics that change are the amount of cache and the "maximum transfer speed", which are specced higher on the SMR drive - but you have no way of knowing that while read may be a bit faster than on the older drive write will be 2x slower.

Again what?  Each drive has a spec'd data rate.   if you buy a drive and the rate is 190MB/s then that is all your guaranteed.  Why would you expect more performance than listed?     Why does it matter if you don't know a drive uses SMR if you are still getting what you paid for? 

 

 

[Kilrah]

Are you purposefully being dense? Go read specs yourself, you'll see all that is specced is a "max speed", conveniently designed to be useless and allowing to hide things like they did here. 

 

[jagdtigger]

4 minutes ago, mr moose said:

Again what?  Each drive has a spec'd data rate.   if you buy a drive and the rate is 190MB/s then that is all your guaranteed.

Sequential speed means nothing and says nothing about the drive's performance in a raid array. Top it off you cant find any mention that the drives use SMR. Stop being ignorant.

[mr moose]

1 minute ago, Kilrah said:

Are you purposefully being dense? Go read specs yourself, you'll see all that is specced is a "max speed", conveniently designed to be useless and allowing to hide things like they did here. 

 

Fuck off with your insults.   

 

Max speed is listed, that is all you are guaranteed.  

[Kilrah]

Just now, mr moose said:

Max speed is listed, that is all you are guaranteed.  

And it's useless to actually judge whether a drive is fit for purpose, and now that it's been used to screw people over they complain. Perfectly normal.

[mr moose]

1 minute ago, jagdtigger said:

Sequential speed means nothing and says nothing about the drive's performance in a raid array. Top it off you cant find any mention that the drives use SMR. Stop being ignorant.

I didn't call it sequential speed, and you think I am ignorant. 

 

 

They don't mention SMR, so what? they also don't tell you what material the platter is made from, they don;t tell the strength of the magnets or material type either.  There is a lot they don;t waste time filling in spec sheets with because for the most part it is pointless information that doesn't have any real bearing on the product.  They give you the maximum data transfer speed.  that's it.  Take it or leave it. 

[mr moose]

1 minute ago, Kilrah said:

And it's useless to actually judge whether a drive is fit for purpose, and now that it's been used to screw people over they complain. Perfectly normal.

 

Since when is the drive I am talking about never been fit for the purpose?  Are you even reading what I have posted?

[jagdtigger]

1 minute ago, mr moose said:

I didn't call it sequential speed, and you think I am ignorant. 

What else can it be on mechanical?   Maybe i am not an IT pro but even i know that 100+MB/s cant be anything else than sequential on spinning rust....

[Kilrah]

4 minutes ago, mr moose said:

Since when is the drive I am talking about never been fit for the purpose?

Read the articles. You might not care if your drive is 2x slower and if it gets kicked of a RAID array if you don't use it in a RAID array, but people who buy drives "specially made for RAID arrays" and the drives don't work in a RAID array do consider it unfit for purpose. 

 

https://blocksandfiles.com/2020/04/15/shingled-drives-have-non-shingled-zones-for-caching-writes/

 

Quote

 I _was_ able to force the WD40EFAX to resilver – by switching off write caching and lookahead. This dropped the drive’s write speed to less than 6MB/sec and the resilvering took 8 days instead of the more usual 24 hours. More worryingly, once added, a ZFS SCRUB (RAID integrity check) has yet to successfully complete without that drive producing checksum errors, even after 5 days of trying.

I could afford to try that test because RAIDZ3 gives me 3 parity stripes, but it’s clear the drive is going to have to come out and the 3 WD Reds returned as unfit for the purpose for which they are marketed.

If you don't give a shit then move on. Other people do and are not going to let themselves being screwed over.

[mr moose]

 

11 minutes ago, jagdtigger said:

What else can it be on mechanical?   Maybe i am not an IT pro but even i know that 100+MB/s cant be anything else than sequential on spinning rust....

Max transfer speed.   That is the best you are guaranteed and that is the only speed spec you are given.  therefore knowing  whether a drive has SMR has no bearing on the performance you are guaranteed. 

 

 

11 minutes ago, Kilrah said:

Read the articles. You might not care if your drive is 2x slower and if it gets kicked of a RAID array if you don't use it in a RAID array, but people who buy drives "specially made for RAID arrays" and the drives don't work in a RAID array do consider it unfit for purpose. 

 

https://blocksandfiles.com/2020/04/15/shingled-drives-have-non-shingled-zones-for-caching-writes/

 

If you don't give a shit then move on. Other people do and are not going to let themselves being screwed over.

You quoted me and started this,  It seems you have ignored some of my posts so you can have a whinge about something else.

 

 

So, as was my point that you seem to have missed in an effort to be affronted,    when you buy a video card, do you look up benchmarks for that model or do you try to gauge how well it will perform based on what other cards do?  Why is a hard drive any different? if you want to know how the ST2000MD001 performs you look up a benchmark with a speed test.  if you want to know how the ST2000DM008 performs you do the same.  Don't buy the ST2000DM008 and then complain you have been ripped of because it doesn't perform the same way as the ST2000MD001.   Seagate never promised you non-SMR performance, you don't get to complain they should give non-SMR performance.   It is really a case that in the domestic market they cannot give you an exact speed, they can't even give you a average speed (all systems are different), Hell they can't even guarantee a minimum speed. 

[jagdtigger]

2 minutes ago, mr moose said:

Max transfer speed.   That is the best you are guaranteed and that is the only speed spec you are given.  therefore knowing  whether a drive has SMR has no bearing on the performance you are guaranteed. 

http://products.wdc.com/library/SpecSheet/ENG/2879-800002.pdf

Hmmm..... No guarantied minimum they only specify "Interface Transfer Rate² up to", which means even less than the sequential. Oh and let me know if you found any mention of SMR.

[Kilrah]

There were no benchmarks to refer to because nobody would expect a respectable manufacturer to do such a dick move. Now there are so people will be able to know, along with a large amount of bad press becasue of said dick move.

 

Until now SMR drives have been specifically designated as such, their limitations pointed out, and they've been marketed only for archival use due to their lack of performance for other tasks. That was honest and perfectly fine.

There was no reason to start sneaking SMR drives into other product lines, somehow completely forget about their drawbacks and consider that something that was clearly communicated about doesn't need to be anymore.

 

Don't understand why anyone would think this is OK. Even less so when WD tried to deny their actions and outright lie.

[mr moose]

3 minutes ago, jagdtigger said:

http://products.wdc.com/library/SpecSheet/ENG/2879-800002.pdf

Hmmm..... No guarantied minimum they only specify "Interface Transfer Rate² up to", which means even less than the sequential. Oh and let me know if you found any mention of SMR.

??

9 minutes ago, mr moose said:

So, as was my point that you seem to have missed in an effort to be affronted,    when you buy a video card, do you look up benchmarks for that model or do you try to gauge how well it will perform based on what other cards do?  Why is a hard drive any different? if you want to know how the ST2000MD001 performs you look up a benchmark with a speed test.  if you want to know how the ST2000DM008 performs you do the same.  Don't buy the ST2000DM008 and then complain you have been ripped of because it doesn't perform the same way as the ST2000MD001.   Seagate never promised you non-SMR performance, you don't get to complain they should give non-SMR performance.   It is really a case that in the domestic market they cannot give you an exact speed, they can't even give you a average speed (all systems are different), Hell they can't even guarantee a minimum speed. 

 

1 minute ago, Kilrah said:

There were no benchmarks to refer to because nobody would expect a respectable manufacturer to do such a dick move. Now there are so people will be able to know, along with a large amount of bad press becasue of said dick move.

Review then,  so if no one does reviews or tests how do you know there is a speed difference?  

 

1 minute ago, Kilrah said:

Until now SMR drives have been specifically designated as such, their limitations pointed out, and they've been marketed only for archival use due to their lack of performance for other tasks. That was honest and perfectly fine.

There was no reason to start sneaking SMR drives into other product lines, somehow completely forget about their drawbacks and consider that something that was clearly communicated about doesn't need to be anymore.

Seagate have never mentioned it on the domestic line up.   

 

1 minute ago, Kilrah said:

Don't understand why anyone would think this is OK. Even less so when WD tried to deny their actions and outright lie.

 

Again, it does not matter for the domestic drives, they were never guaranteed to have specific performance, they were never listed as having or not having it.  For majority of the uses an MD drive sold as being for,  the end user will rarely hit write speeds that would be notably hampered by SMR.   Most writing speeds are limited by the source (ODD or internet) not the technology used int he drive itself.  

[Kilrah]

5 minutes ago, mr moose said:

Seagate have never mentioned it on the domestic line up.   

Of course not since the domestic lineups never included SMR drives before... Their use was limited to those archival drives, alledgedly becasue they were considered too poor to even go in domestic drives and their use required disclosure of that feature. The problem is that they decided to silently change this stance...

[jagdtigger]

2 minutes ago, mr moose said:

??

 

24 minutes ago, mr moose said:

Max transfer speed.   That is the best you are guaranteed and that is the only speed spec you are given.

Its only for the interface and not even guarantied:

Quote

² As used for storage capacity, one megabyte (MB) = one million bytes, one gigabyte (GB) = one billion bytes, and one terabyte (TB) = one trillion bytes. Total accessible capacity varies depending on operating environment. As used for buffer or cache, one megabyte (MB) = 1,048,576 bytes. As used for transfer rate or interface, megabyte per second (MB/s) = one million bytes per second, and gigabit per second (Gb/s) = one billion bits per second. Effective maximum SATA 6 Gb/s transfer rate calculated according to the Serial ATA specification published by the SATA-IO organization as of the date of this specification sheet. Visit www.sata-io.org for details. Performance will vary depending on your hardware and software components and configurations

Oh yeah, and they use decimal for the transfer rate and not binary.....

[mr moose]

Just now, Kilrah said:

Of course not since the domestic lineups never included SMR drives before... Their use was limited to those archival drives, alledgedly becasue they were considered too poor to even go in domestic drives. The problem is that they decided to change this stance...

Yes, they started putting SMR into domestic drives and didn;t tell anyone why? because it wasn't an earth shattering change to how a drive is used.  Either no one noticed (including me) or it wasn't the world ending problem you are making out.  You see people are still getting the the drive they paid for, in fact they are getting a bigger drive because of SMR.  So it really is the perception of missing out on performance,  not that the you were actually shown or promised that performance in the first place.

[mr moose]

33 minutes ago, jagdtigger said:

 

Its only for the interface and not even guarantied:

Oh yeah, and they use decimal for the transfer rate and not binary.....

Max transfer speed is the slowest of the total internal and external data transfer speed of a drive.

 

It is not for the interface only.

 

https://www.seagate.com/www-content/product-content/barracuda-fam/barracuda-new/en-us/docs/100805918l.pdf

 

if you look at table 2.1 there are two figures, one for the interface and one for the drive as a whole.  the interface is 600MB/s (sata standard) the other is the max speed at 190MB/s. 

 

EDIT: just to add to this, the drive itself won't even saturate an original  1.5Gb/s sata connection.

[jagdtigger]

53 minutes ago, mr moose said:

Max transfer speed is the slowest of the total internal and external data transfer speed of a drive.

WD datasheet only specifies interface speed, so stop claiming things that are not there(nice try to move the goalpost BTW). And its not even guarantied as its mentioned in the fine print.

[mr moose]

5 minutes ago, jagdtigger said:

WD datasheet only specifies interface speed, so stop claiming things that are not there(nice try to move the goalpost BTW). And its not even guarantied as its mentioned in the fine print.

For the love of everything. please read my posts before making making accusations. I have never refereed to WD, I don't give a fuck what they say on their page.

 

Why would you do that? why,  when I talk about a seagate specs and have been linking only to it since midway through the first page would you go to a wd page to try and prove me wrong? what logic?

 

EDIT: also the generic definition for maximum transfer is the drive and the interface according ti wiki, And unless you were born under a rock you'd know that 190Mb/s is fucking slow for a sata interface only.  so unless every HD maker is still only using sata 1 (oh no they are not), then trying to refute what I said by claiming it's only the interface is nonsensical at best.

[jagdtigger]

8 minutes ago, mr moose said:

For the love of everything. please read my posts before making making accusations. I have never refereed to WD, I don't give a fuck what they say on their page.

 

Why would you do that? why,  when I talk about a seagate specs and have been linking only to it since midway through the first page would you go to a wd page to try and prove me wrong? what logic?

Yeah fine, whatever. Even if we take seagate's data it wont matter anything because its a very edge case where the drive can spout out that much data, which you will never reach in real life. Its totally not indicative of anything just like in case of SSD's.

[mr moose]

4 minutes ago, jagdtigger said:

Yeah fine, whatever. Even if we take seagate's data it wont matter anything because its a very edge case where the drive can spout out that much data, which you will never reach in real life. Its totally not indicative of anything just like in case of SSD's.

Which is my whole point,  drive speeds fluctuate between 80MB/s and 180MB/s.  They change so much depending on the drive, the system, the usage and even the temperature.  Trying to claim SMR has halved the performance* of a domestic drive when we don't even have a uniform position on what it should do is crazy,   let alone the idea that not knowing if the drive having SMR even plays a role in that.

 

*which is exactly what Kilrah tried to argue.

[Sauron]

11 hours ago, Curufinwe_wins said:

Part of the reason why these drives are causing RAID issues is because even in software/firmware, they are not properly reporting their technology. So the array thinks something is totally boned and kicks out a failure.

 

This is a serious issue. On multiple levels.

That's really interesting - I would say this is probably a bigger problem than the drives being slower. If they reported their technology properly the array could just wait for them... I wonder what possessed these manufacturers to think this was a good idea. If they are good enough for NAS and RAID (which they probably are) then why not just own the fact that they are different?

8 hours ago, davide30541 said:

but is this not a software issue? its an hdd that works, only in a way that nas OS are not expecting. they treat these new hdds as broken because they behave differently from the ones in there already. once nas OS is aware that these drives are diferent this will become a non issue? just thinking here. 

A drive behaving differently than what it claims to is usually a clear sign of failure. The OS can't just go and guess what technology the HDD has if it's not told.

[Curious Pineapple]

Great news! I can finally start selling batteries with "up to" 20000 recharge cycles, then in 5 years I can increase the capacity at the cost of recharge cycles but as long as I say "up to" people will defend me!

[mr moose]

13 minutes ago, Curious Pineapple said:

Great news! I can finally start selling batteries with "up to" 20000 recharge cycles, then in 5 years I can increase the capacity at the cost of recharge cycles but as long as I say "up to" people will defend me!

 

I get the hatred for the term "upto", but that really isn't the cause of the issue here.  Nor is it blanket get out of jail free card.    There is a specific problem with products of this nature in that manufacturers can't actually tell you how fast the drive will work in your end use conditions, only give you a best case scenario.   Otherwise how do they work out a speed that is actually meaningful to a buyer?  how do they not spec it to look slow when it can perform faster but don't want to invoke legal issues.  you know why we have a bakers dozen right?, it's the same thing with performance metrics.  an "Upto" figure doesn't land them in consumer law issues while any other figure might.   The best they could do is measure ever one's real world performance and advertise it as an average.  But you know how well that would go down.