I bought the FuzeDrive Plus now, even if the “free” version was not working because of Installer incompatibility.
In fact the free version would have been inappropriate for my system, I would have had to split my NVME.
So now I am waiting for the drives to be entiered and then we will see how good it works…
and it’s finished allready, I don’t think I will really feel the difference in speed, my drives are allready very fast, but I gained more space, that’s nice.
with drives/setup that large, give it about 2-3 weeks to accumulate use counters on the blocks/data. with the 280gb optane + 1tb regular ssd… it made quite the difference.
SSHD is crap because they include very little flash and aren’t very intelligent about utiisation of it.
What would be nice is if you could tag files with some sort of metadata tag as “archive” or something like that and have them never moved into the fast tier.
If having a drive failure causes you to lose data you are doing this whole computing thing wrong.
Have backups.
Repeat after me.
Have backups.
Ideally with a copy off-site.
Drives fail.
PCs fry components
PCs get stolen
User error happens
Malware happens
OS bugs happen
Application bugs happen
Houses burn down
etc.
Do not ever rely on a single copy of your data. If i can go into your house and cause you to lose a significant amount of data that you can’t get back, you are doing it wrong.
Thanks for the condescension, friend. I do take backups. I don’t want a drive failure to force me to restore them, because that’s a huge pain in the ass.
Yep. Come to think of it, that’s actually an area where ZFS, being a file system, could beat the shit out of traditional tiered storage, if they ever add support for more than just log/cache tiers.
Well if you’re not doing RAID and you have a single drive failure of the spinning drive, you’re boned in that case as well.
Very little difference, and certainly not really a reason to poo on tiered storage as a performance enhancer. Anecdotally, these days you’re more likely to see a failure in the spinning rust drive anyway, and storage review’s actual SSD torture test study tends to back that assertion up.
Of course there’s a difference, with tiered storage there are twice as many points of failure.
The primary difference between tiered and caching is that with tiered storage you benefit from the SSD’s size and end up with larger volumes. But typically you use a small SSD and a very large magnetic drive to do this, so your resulting volume might be 6TB rather than 6.25TB, which is not a big deal.
If you like that tradeoff go for it. I wasn’t saying it has no place, just that I wasn’t comfortable using it myself.
It seems that AMD removed the X399/sTR4 language from their FAQ and this isn’t available on Threadripper anymore? Any idea what happened there?
@wendell
How does this differ from Microsoft Tiered Storage Spaces? Other than StoreMI seems to actually work.
I have two 120GB SSD’s and two 6TB HDD’s.
I don’t get it, performance of my tiered storage space on Server 2016 is worse than the HDDs’ in raid1. Like much worse, sequential writes in raid 1 are 230 where as with the tiered storage space, I’m getting erratic speeds from zero to 34, then jumping to 100 and back down to zero. Weirdly, when I run a crystal disk speed test it works pretty well. But as soon as I begin a large file copy (robocopy my files back over from an older single disk) it sux and speeds drop lower than the software HDD mirror.
Any help or suggestions would be appreciated. At the end of the day, its a NAS and the HDD’s are fast enough for the network, but it’s really annoying me I can’t seem to get it to work. Plus I want to use those SSD’s.
Here are my PS commands after I get the drives installed:
New-StorageTier -StoragePoolFriendlyName “Mirror Pond” -FriendlyName SSD_Tier -MediaType SSD -ResiliencySettingName Mirror -NumberOfColumns 1 -FaultDomainAwareness PhysicalDisk
New-StorageTier -StoragePoolFriendlyName “Mirror Pond” -FriendlyName HDD_Tier -MediaType HDD -ResiliencySettingName Mirror -NumberOfColumns 1 -FaultDomainAwareness PhysicalDisk
$ssd_tier = Get-StorageTier -FriendlyName SSD_Tier
$hdd_tier = Get-StorageTier -FriendlyName HDD_Tier
New-VirtualDisk -StoragePoolFriendlyName “Mirror Pond” -FriendlyName Vault -StorageTiers @($ssd_tier, $hdd_tier) -StorageTierSizes @(22GB, 5588GB) -ResiliencySettingName Mirror -WriteCacheSize 95GB
Set-StoragePool -FriendlyName “Mirror Pond” -IsPowerProtected $True
Then I format the Virtual disk to ReFS and restart
So can I use say Optane 800p as swap/Virtual Memory yet? :3
Obvious, but just sayin, if u ssd exceeds the 128/256GB fast tier limit, it allocates the remaining space to a separate drive letter.
So as a way of a/ increasing fast tier size & b/ fixing files to the fast ssd, …
Windows page file and temp files e.g could be configured to reside on the spare ssd space, resulting in; more tier 1 cache space, less writing to and fro between tiers & same performance.
Great video, @wendell. I’m looking forward to your video about tiered storage on Level1Linux.
Ideally, I’d like a local write-back cache on Optane for a NAS volume. This would give minimum latency for reads and durable writes on my workstation, while the NAS takes care of redundancy, backups and provides a large amount of storage. Does anyone know of a way to do this on Linux or Windows?
I realise I’m quite late to the party but I just watched this video.
What’s the difference between Tiered Storage and PrimoCache (HDD, SSD RAM), seems only the sizes involved?
Does anyone still use compression on their storage media, like LZ4
I haven’t used anything since Stacker, bad experience, is there anything worth using on Windows10?