Intenso SSDs I read VERY BAD Things about, I assume that's why you have triple Redundancy.
About Spinning Rust no complains so far with HGST / Toshiba, but knocking on Wood ...
Proxmox with Second Hand Enterprise SSDs
- Thread starter silverstone
- Start date
About Spinning Rust no complains so far with HGST / Toshiba, but knocking on Wood ...
heh, I just did the same. 2x 1.6TB for $130AUD each. Have only just loaded them up with Proxmox. One shows 1% wearout the other shows 2%. Both with 50,000 hours (not really a concern for SSD - I have had plenty of consumer SSD make it to 10 years power on without issue, which makes sense for SSD). But yeah, I was taken aback with the extra smart values shown for these drives. For example, this is my Samsung 893 (I bought new):
And these are the two Intel used drives I bought:
Reminds me of when you look at a WD HDD and then go look at a Seagate with all its extra and different values.
Alex, how thrashed do the used ones I bought look like from the above? I guess I should just be happy they're 1 and 2% wearout. Planning on either passing them through to UnRaid VM as cache drives, VM storage drives, or to be used with Usenet for initial downloads before being moved to spinning rust.
Either way, I was personally willing to take a chance on some ebay used drives after shelling out almost $300AUD for a single new Samsung 893 960GB
Edit: Just passed one through to UnRaid and laughed at the Total lbas read/written count:
Last edited:
I have no complaints about any brand of spinning rust I have used and that includes... get ready for it... are you ready?? Seagate. Have WD Reds upwards of 10 years old, have Seagate Barracuda Pro and Exos doing great (Exos drives have 5 years warranty, also). Everything is sweet *touches wood*. It's funny when I see people swear off entire companies because of a bad batch or experience. I remember in 2009 Seagate had a shocker with the 7200.11 desktop drives. But I couldn't turn down the value for money of shucked Barracuda Pros and the beloved noisy but reliable Exos drives. People even start harping on about backblaze reports - not even noticing that Seagate is the biggest riser in reliability now that the 4TB Seagates are going out of service.
sdn looks fine; sdm and sdj have >55000 poh... I'd keep backups.
power on hours should be mostly irrelevant for ssds as there are no moving parts that could degrade.
keeping backups is always a good advice though
keeping backups is always a good advice though
Yeah, that's pretty much how I'm approaching it. I bought these used enterprise SSDs just last week, so they came with the 6 years power on. I've had SSDs last over 10 years easily, and their downfall wasn't from the power on hours, it was from the amount of data written to them in the end. For $130AUD for 1.6TB enterprise SSD, I was willing to take the risk with 50k+ hours on them. Either way, my plan is to use them for Usenet download drives, before the completed files get moved to the main HDD ZFS arrays. They won't be holding anything that cannot be replaced. But from what I've read, they should have years left in them.
this is my personal opinion, but i dont see any issue with the high power on hours.
there is a small chance that the capacitors on the drive have aged and dont have their designed capacity anymore.
i have no personal experience with any ssd experiencing this, but since capacitors do age and can fail, there is always a possibility for it.
newer drives such as the s4510 have a smart value indicating the health of the plp capacitors:
unfortunately the s3610 doesnt have that. at least my dell branded version doesnt show it in smart.
maybe someone with the intel branded version can check?
there is a small chance that the capacitors on the drive have aged and dont have their designed capacity anymore.
i have no personal experience with any ssd experiencing this, but since capacitors do age and can fail, there is always a possibility for it.
newer drives such as the s4510 have a smart value indicating the health of the plp capacitors:
175 Power_Loss_Cap_Test 0x0033 100 100 010 Pre-fail Always - 2659 (21 65535)235 Power_Loss_Cap_Test 0x0033 100 100 010 Pre-fail Always - 2659 (21 65535)unfortunately the s3610 doesnt have that. at least my dell branded version doesnt show it in smart.
maybe someone with the intel branded version can check?
Last edited:
device age isnt relative to moving parts (at least not exclusively.) nand have known limits in their behavior; even if they havent exceeded their TBW limits doesnt mean you dont keep pulling the trigger on the proverbial russian roulette gun the longer they are in operation.
Ultimately none of this really matters. I would not deploy such drives "for real" but for a homelab- sure! the consequences of failure/poor performance arent going to put me out of business/lose my job.
I'm a marine scientist, so I only discuss and care about my homelab, lol. Obviously, in a business setting you'd be going for new drives. I should have prefaced that when posting my initial reply with my drives. I forget people here do this stuff for a living.