Even 4 bay hardware RAID enclosures seem a little sparse to be honest. 2 bays are available, but felt a little restrictive for future expansion. Mediasonic were the only ones I could really find, but they split their 4 bays into 2 pairs of RAID setups, so can't RAID across all 4 drives, e.g. RAID0 across the entire enclosure as one set of 4 drives. But even so, Mediasonic are not available in the UK/Europe and Raidsonic products seem to be sold under the IcyBox branding, but with significantly more limited options that the US/Far East, particularly within the UK.

I've taken a punt on the TerraMaster D6-320 (which came from left field a bit and not previously considered). Not sure if I will keep or return yet though.

I was contemplating using is as a single power/USB cable, it's 10Gbps USB 3.2 Gen 2, and then use either Windows Storage Space or StableBit Drivepool to manage the pools and duplications between the drives, containing it all within the one enclosure. It also has 2 free bays that I can hot swap in other drives if need be then too.

Not sure how successful it will be, but it should be arriving today and the IronWolf Pro drives are already here waiting for it..

 

Hardware RAID is not "dead" despite what some you-tuber(s) might think. You should consider the source on that one... For mission critical/high reliability applications and database in IT and DC's etc, you wont see software based RAID running in these scenarios, and there's a reason for it. It may be overkill in lower key home situations such as what you're describing here but that doesn't make it "dead" technology.

 

Last edited: May 31, 2024

I did take the video with a pinch of salt. But then did more research and found that more and more consumer stuff (perhaps not professional enterprise grade stuff) are moving away and hardware RAID offerings are becoming more and more sparse, be it QNAP, Synology, OWC, Asustor, TerraMaster, Sabrent, Raidsonic, and so on, they are all being released with less and less hardware RAID options it seems.

My use case for these drives and external enclosure is also not necessarily for mission critical, so less of a concern. My process (I hope) is:

DESKTOP: 10TB of NVMe and 8TB SSDS overall, to include mission critical data and all to be backed up to Drive Pool 1

DRIVE POOL 1: To be set up in RAID0 (or equivalent software pool) and containing entire Desktop backup as well as other non-mission critical and less regularly accessed/required data, including archival content. All to be backed up to Drive Pool 2 and Cloud Backup

DRIVE POOL 2: To be set up in RAID0 (or equivalent software pool) and be a mirror of Drive Pool 1 and serve as local backup for Drive Pool 1

CLOUD BACKUP: To be a full back of the Drive Pool 1 (and therefore also Desktop)

REDUNDANCY: Though Drive Pool 1 and Drive Pool 2 are set up as RAID0 (or equivalent software pool), they provide RAID10-like redundancy for each other and due to using a software drive pool, there is no rebuild or down time required after a drive failure, as it is a case of copying the content from the working Drive Pool 1/2 drives onto a replacement drive. Enclosure is hot swappable, so replacement drives can be added easily.

FUTURE EXPANSION: The enclosure has 2 empty bays to use as hot swap drive for spare drive in the short term, but also serve as expansion for further 18TB in the future to expand Drive Pool 1 and Drive Pool 2 when required.

ENCLOSURE FAILURE: As both Drive Pool 1 and 2 are within a single enclosure, there is a risk of enclosure failure. The pools are not enclosure dependant, and therefore the drives can be moved into a new single or separate enclosures if need be and should still be recognised by Storage Spaces/Drivepool. If this change of enclosure (either to an indentical or different enclosure(s)) fails, then the Fail Safe would be needed.

FUTURE OS CHANGE/REINSTALL: Windows Storage Spaces and Drivepool both write pool setup and identification data to each drive, meaning they can be read as pools by other computers. This means future Windows upgrades will not lose the pool (unless Microsoft drop Storage Spaces or Drivepool is end of life'd). If Storage Spaces or Drivepool is dropped, I can call on the Fail Safe if needed.

FAIL SAFE: The fail safe is that everything is also backed up to the Cloud Backup, meaning if there is a catastrophic failure, e.g. multiple drives failing at once or becoming corrupted, enclosure change issue, and so on, the data can be re-downloaded onto new drives from the Cloud Backup

The above makes sense (I think) in providing me with 3 local copies of mission critical data (Desktop, Drive Pool 1, and Drive Pool 2) and an additional 1 off-site copy (Cloud Backup), 2 copies of local less critical data (Drive Pool 1 and Drive Pool 2 with 1 off-site copy (Cloud Backup), and provide an absolute fail safe for any worst case scenario (with the exception of Backblaze data being compromised, which I guess you can't really presume nor account for realistically.

However, for me to lose everything, I would need my Desktop to fail, both Drive Pools to fail, and Backblaze data to be compromised at the same time, which the risk of is probably tiny in the grand scheme of trying to ensure there is an appropriate back up process for data.

 

Last edited: May 31, 2024

In a way, it's funny how at this rate you will end up with a backup strategy that has more contingency levels than whatever my local administration came up with for storing life-and-death critically important data.

And yeah, software RAIDs running on low powered ARM, MIPS, RISC, PPC chips are cheaper to realise than designing RAID chips + proprietary software stack.

 

Most CT-scans are originally stored on a raid, as its a fast way to get the data out of the imaging system where it can be manipulated with 3d software and reconstruct the scan into an image a doctor can read. Siemens,Philips,Ge,Toshiba have been building raids into Ct scanner operator consoles for decades.

Its not dead technology, its dead perfect technology for capturing huge amounts of data fast and safely.

 

debate of hardware vs software RAID really comes down to trust of technology and company,
hardware RAID by design is a proprietary tech, when it dies (or company ceases to exist), you may be lucky to find same (or compatible) hardware RAID to move drives into,
software RAID may use various open-source elements, so when it does, there's a higher chance of still being able to access your data, but again such software component or company/support may cease to exist,
for ex. ZFS pools can be migrated between various systems that support it (for ex. Proxmox and TrueNas) or you may migrate Synology into Xpenology (and vice versa)...

...but, always have proper backups - and RAID is NOT a backup, it's merely a convenience of zero-downtime being able to hot swap dead drive(s) without interrupting operation and data access

 

Storage media life expectency. HDD, SSD,optical etc.