You wrote: "What is the current thinking regarding SSD replacement for rotating storage in the ReadyNAS system?"
That heavily depends on the major kind of purpose your nas'es are operated for. If one focuses on streaming or archiving hdd are fine, imho. If you need fast random access instead, ssd might be the first choice.
However, the costs of ssd compared to hdd of the same size might keep you from going for ssd, anyways.
You wrote: "What is the current thinking for eliminating Raid 1 two disks versus one SSD and then having expanded capacity?"
That depends on the number of good backups you might have available. There is a golden rule: If you have 3+ instances of your file available in three different places you could start considering to be "safe".
I also have a question to you. You wrote "The disks are getting 'long in the tooth' and a smart person would start to replace."
What is the indicator you use for your judgement to replace disks. Are those "just" old or are there any errors reported in the logs?
Our ReadyNAS systems are deployed in small remote offices -- essentially semi-autonomous locations with no technically skilled personnel.
Access is for business type applications – network traffic is modest at best.
My concern is focused on the cost of correction – time the systems are down and personnel available to fix the issues plus lost productivity. That said, the actual additional cost of SSD units over the expected life expectancy of five plus years becomes inconsequential to the other factors.
I have never heard of the 3+ golden rule but you answered my question from your view – keep your current Raid1 approach for the next round of disk replacements.
My notions of disk replacement relates to their current age and not any measured error rates. To my knowledge we have had only an occasional error circumstance. We have almost 50 of these systems in place and the cost of monitoring them remotely drives me to do replacements as a prophylactic measure not as a reactionary measure.
What is the current thinking regarding SSD replacement for rotating storage in the ReadyNAS system?
What is the current thinking for eliminating Raid 1 two disks versus one SSD and then having expanded capacity?
I can only give you my own thinking (and I still have mechanical disks).
I haven't seen any data that says that SSDs are more reliable than mechanical disks. SSDs certainly will give you faster random access - browsing large folders, database access, etc. They generally won't help much with sequential access, as that is limited by network speed, etc. They also use somewhat less power.
On RAID-1, I'd still use it. But I'd stagger the drive replacement schedule - scheduling a replacement of one drive in each pair about half-way through it's expected life. My thinking there is that SSDs have a limited number of writes - and if the disks are mirrored, both drives will reach that limit at the same time. So if the replacements are staggered, you reduce the risk of simultaneous failures.
OS-6 systems have a middle ground btw - you can use SSDs for metadata tiering (and in the next release for data caching). That's called "ReadyTier".
Any other issues or concerns that will byte us in the A##?
AFAIK the NAS web ui won't give you any indication of the total bytes written (or the wear level reported in the SMART stats). You could monitor that by downloading logs, or you could write your own app/script which reports that information.
Also, OS-6 NAS drive trays support 2.5" SSDs. But older NAS drive trays don't. You'd have to use an SSD drive bracket to put the drive in the correct position. That bracket might need to be glued to the trays somehow.
Some of the maintenance functions ( scrub, balance, and perhaps even defrag) do a lot of disk writes. So you'd want to trade off the benefits against the reduction in drive life.
NAS backup comes up pretty frequently here, as many people make the wrong assumption that RAID is enough to keep their data safe.
I think @Retired_Member is referring to my own backup policy - which is to have at least 3 copies of everything I care about on different devices (including the original). That reduces the risk of losing data when the primary device fails ( and then you learn that the backup is corrupted somehow). Plus if one copy is off-site, and another is on-site you can recover quickly when needed, but still have protection from a disaster.
