Western Digital HGST Ultrastar Hs14 Enterprise 14TB Hard Drive Launched

[Cliff Robinson]

Western Digital HGST Ultrastar Hs14 Enterprise 14TB helium-filled hard drives are launched with host-managed SMR and both SATA III and SAS 3 interfaces

The post Western Digital HGST Ultrastar Hs14 Enterprise 14TB Hard Drive Launched appeared first on ServeTheHome.

Continue reading...

 

[Evan]

Ok so out 9 months from announcement to launch...
I actually don’t really see why you would use 14tb SMR over 12tb PMR unless there is a substantial cost reason.

How large can we go

HGST is at 8 platters now in the He12 line so I don’t see any chance at all to add more, only density increase... we are getting more or less to end of the line until he get some newer tech like heat assisted going. Meanwhile 8tb is rather large and reasonably cheap.

 

[K D]

Larger data centers and systems providers want to manage the SMR process which adds a twist.

From what I know this is internal to the Hard drive and is managed by the drive's internal controller. What does this mean? Are there systems that currently do this?

 

[Evan]

From what I know this is internal to the Hard drive and is managed by the drive's internal controller. What does this mean? Are there systems that currently do this?

3 types of SMR drives...
- drive managed; all logic handled in the controller (think retail archive drives)
- host aware ; some logichandled by the host, this means you can use some software on the host side to assist or optimise with how the drive writes the data.
- host managed ; needs host software, extensions to ata/scsi command set required

I Guess theses drives may be host managed or host aware versions, the hyperscalers use host managed sometimes as they have control of the software. Those host managed disks won’t work with regular controllers etc.

So yes systems that do that, linux kernel extensions, file system changes etc.
Never seem or used any SMR drives actually, just not an area have any need to play in but I am sure somebody lurking here it may be their daily job who has more info.

 

[realtomatoes]

3 types of SMR drives...
- drive managed; all logic handled in the controller (think retail archive drives)
- host aware ; some logichandled by the host, this means you can use some software on the host side to assist or optimise with how the drive writes the data.
- host managed ; needs host software, extensions to ata/scsi command set required

I Guess theses drives may be host managed or host aware versions, the hyperscalers use host managed sometimes as they have control of the software. Those host managed disks won’t work with regular controllers etc.

So yes systems that do that, linux kernel extensions, file system changes etc.
Never seem or used any SMR drives actually, just not an area have any need to play in but I am sure somebody lurking here it may be their daily job who has more info.

Would that mean when these hit the used market, these drives can’t be used outside of the manufacturer’s approved hardware?


Sent from my iPad using Tapatalk

 

[cperalt1]

Correct they require Host managed SMR so unless there is an open source implementation then they will be locked to original system.

 

[Evan]

Would that mean when these hit the used market, these drives can’t be used outside of the manufacturer’s approved hardware?


Sent from my iPad using Tapatalk

Software really, and there is Linux related code so I wound think it’s generally open source.
It’s mostly going to be implemented in kernel and file system.

 

[nkw]

Software really, and there is Linux related code so I wound think it’s generally open source.
It’s mostly going to be implemented in kernel and file system.

I think Ceph/Bluestore + HM-SMR is intriguing, but appears to be a ways off. I was researching this last week -- if anyone is interested in the current state of SMR support amongst ext4 and ceph this is what I found:

GitHub - Seagate/SMR_FS-EXT4 -

SMRFFS is an addition to the popular EXT4 to enable support for devices that use the ZBC or ZAC standards. Project scope includes support for Host Aware (HA) devices, may include support for Host Managed (HM) devices​

Evolving Ext4 for Shingled Disks | USENIX -

We introduce ext4-lazy1, a small change to the Linux ext4 file system that significantly improves the throughput in both modes. We present benchmarks on four different drive-managed SMR disks from two vendors, showing that ext4-lazy achieves 1.7-5.4x improvement over ext4 on a metadata-light file server benchmark. On metadata-heavy benchmarks it achieves 2-13x improvement over ext4 on drive-managed SMR disks as well as on conventional disks.​

https://image.slidesharecdn.com/201...for-ceph-one-year-in-44-638.jpg?cb=1490305768 -

Recent work shows less-horrible performance on DM-SMR – Evolving ext4 for shingled disks (FAST’17, Vault’17) – “Keep metadata in journal, avoid random writes” ● Still plan SMR-specific allocator/freelist implementation – Tracking released extents in each zone useless – Need reverse map to identify remaining objects – Clean least-full zones ● Some speculation that this will be good strategy for non-SMR devices too​

Re: bluestore SMR update? — CEPH Filesystem Development -

To me this suggests that bluestore will probably reasonbly well as is on
drive-managed SMR disks for RGW workloads (which write larger objects in
their entirely). Different drives have different approaches to managing
the 'scratch space', though; with a bit of testing we can probably
recommend which ones will work best (my guess is those that demonstrate
fixed overhead instead allowing a large amount of deferred work to
accumulate). ... Longer term, I think it may still be best to have bluestore manage it's
overwrites more carefully, but depending on how well the drive-managed
ones work it may not be worth the investment.​

 

[Bert]

Is there a way to make host managed SMR drives work through Linux distros? I read an article which says in recent kernels btrfs supports it. Did anyone here give a chance?