Rereading articles, it seems that I was incorrect, in a way.

Being that few "regular people" read and write as many times as the SSD's are rated, I guess the SSD do last. HDD are suppose to last forever but I would guess that the mechanical parts will go bad after a very long time.

For those who nearly fill their SSD, that tells me the unused portion will be used lots over and over again. That will wear the unfilled portion out much faster than is a much larger unused portion.

So to me that means either will store well enough for "us regular people", but not is the SSD is nearly filled to begin with.


Here are some quotes from the following articles:


1) https://www.exittechnologies.com/blog/it-tips/hard-drive-vs-ssd-vs-storage-system/

"HDD’s are more vulnerable to mechanical failure since they contain moving parts...
Usually, they will warn the user before failure as they start to make a lot of noise or load data slowly."

"Solid-state drives do not suffer from mechanical problems...
However, the electrical circuits that contain the data can wear out and fail. Writing data is harder on the drive than reading data.
One unique issue that SSDs have that hard drives don’t is electron wear.
When writing a 1 or a 0 to the same location repeatedly, that cell can begin to wear out. This is mitigated through a concept called wear leveling or provisioning.
Normally, an SSD fills up to 60 % and that data doesn’t change much, but the remaining 40% changes often as the user creates and deletes files.
Once a failure occurs in that overused 40%, the whole drive fails. The provisioning function reorganizes the information on the drive as needed. That way all sectors are being worked evenly, prolonging the life of the drive."


2) https://www.howtogeek.com/322856/how-long-do-solid-state-drives-really-last/

"A joint study between Google and the University of Toronto covering drive failure rates on data servers. The study concluded that the physical age of the SSD, rather than the amount or frequency of data written, is the prime determiner in probability of data retention errors. It also determined that SSD drives were replaced at Google data centers far less often than conventional hard drives, at about a one to four ratio. But it wasn’t all positive in favor of SSDs: they experienced higher uncorrectable errors and bad blocks at a much higher rate than hard drives over the four-year testing period. Conclusion: in a high-stress, fast-read environment, SSDs will last longer than hard drives, but be more susceptible to non-catastrophic data errors. Older SSDs are more prone to total failure regardless of TBW or DWPD."

Good information. As an added data point, all modern versions of Windows issues the TRIM command with each write, which accomplishes wear leveling. I can’t speak for Macs.

Btw, I am now on my 6th year with eight Intel 500 series SSDs spread across several machines. I routinely run at least one drive (usually the system drive) >90% capacity, but it, like most consumer machines, doesn’t get nearly the type of access as a commercial server. It’s not good practice, but it seems that many of the aps and system files are constantly growing, and I get behind on reallocating files to other drives.

You have a common misconception about what RAID is for. RAID is not for backing-up or archiving. People erroneously think that having a RAID mirror setup is the same thing as one drive backing up another. It isn't so. That's what external drives would be for. RAID is for multiple drives acting as ONE DRIVE, or redundancy, meaning if a drive in a RAID array fails, it gets replaced and the RAID controller configures the new drive to take its place in the array. A RAID array should be thought of as ONE DRIVE, not drives backing up one another.

A drive out of a RAID array cannot be read by another computer by, say, connecting it via USB. A RAID drive can only be read by the controller (hardware or software) that created it. A proprietary formatting system is used in a RAID system. I learned this when I first set up my NAS. I used two drives as a RAID mirror thinking just as you are...the primary drive fails, just read the second drive on another machine. The drive COULD NOT BE READ on another machine. I re-built my NAS with just extra USB drives, formatted normally, and the NAS operating system syncs the extra drives with the primary.

What you CAN do is have two or more drives in your computer, and use a drive synchronization program to duplicate the primary drive to other internal drives. Or, have just one data drive in the computer, and sync it with as many externals as you like for archival/backup purposes.

In general, I agree (and thanks for pointing this out). One exception is a SW RAID 1 mirror managed by Windows. If both drives are NTFS (for example) and are written by Windows, the DATA should be readable by another Windows machine. That doesn’t mean though, that you can plug the drive into a different Windows machine and boot the OS from it (assuming you mirrored the OS as well).

Another common RAID misconception is that you can remove one drive from RAID set (redundancy group), such as a RAID 5, periodically, store it off-site as a DR copy (occasionally swapping it for a 2nd spare drive which the RAID will then rebuild). Except for RAID 1 (simple mirroring), the removed drive only contains a portion of your data. I have had this point argued to the death by a RAID user who believed that somehow, magically, all the data from the other drives was somehow contained on this single drive, which could then be used to restore all the data in the event of a RAID failure. He was dutifully swapping out one drive and storing it off site (essentially creating a drive failure in the RAID set) and repeating this operation regularly, alternating the drives. Fortunately, he had not had to test this method in the event of a real failure.

An example the opposite of your Windows-managed RAID 1 array experience:

My NAS, based on a Linux OS (OpenMediaVault), configured the RAID drives (software RAID) with a proprietary file system, NOT with a standard Linux file system (ext3 or 4). The resulting drives could not separately be read on a Linux machine. They were only usable in the array on the NAS.

I researched this phenomenon and learned RAID was not the solution for plain-old backing up, which is what the uninformed (like I was) think they are doing.

I rebuilt the NAS with the network-accessed drive in Linux format (which OMV recommends), and multiple USB-connected drives formatted as NTFS, which are the backups that are synced with the main drive. The backup drives can be read on any computer...Windows, Linux, or Mac.

Just to clarify. Let’s not confuse anyone with the two separate issues - the file system (NTFS in this case) and the manner in which a RAID system lays down data. The data on an NTFS formatted drive (or a FAT32 or...) can be read by any platform that supports that file system IF it is a plain, unstriped and unencoded file. The individual files can also be read in the same way IF they were written as a true mirror (RAID 1) of the original.

NAS and RAID arrays often use a proprietary file system, so only a machine using the same file system can read them. Secondly, since the data is byte or block striped as is true of other RAID levels, the RAID controller managing the array must use the same algorithm (and settings such as block size) as the controller that wrote the data. The net-net is that for levels other than RAID 1, the same type of RAID controller, using the same algorithm, settings and RAID level, must be used to read and reconstruct the original data.

Bottom line: simple mirroring (RAID1) is an exception to the above IF and ONLY IF the file system of the written data is capable of being read by the system doing the reads. NTFS is a relatively universal FS in that it can generally be read by Windows, Macs and Linux platforms.

Not sure what you are clarifying.

1. My RAID 1 system did not use NTFS. It used a proprietary format. Okay, some RAID 1 systems may use NTFS, but this is definitely NOT a standard practice. And that is the whole issue with attempting to use RAID 1 as a backup strategy.

2. Your bottom line is no different than what has already been said.

We aren't in disagreement here. My whole point is that for the person who maybe doesn't understand all this, and doesn't need to, the big takeaway is.....avoid RAID as a simple backup method. Use all drives with NTFS (with Windows), and backup the primary drive to others using a syncing program.