New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Hetzner SSD failures likely?
Hi,
I'm debating getting a Hetzner dedicated server w/ datacenter SSD from their server auction. Is there anything sensible that can be said about the likeliness of an SSD failure (and the reason to get mirrored SSDs rather than a single one)? I'm using Proxmox so it's trivial to backup the rpool anyway with a snapshot (both to a local HDD and a remote backup).
Any suggestions / tips?
Comments
I'd suggest just checking the drives when you get the server, running some health checks, and making evaluations.
https://usercontent1.hubstatic.com/13054240.jpg
Courtesy of Backblaze, Drives are more prone or likely to fail in the early beginning of their life, and consequently at the end. If the drive is somewhere between there close to the early that late, you're probably save.
These disks ends up being loads better than my Nocix disks which had power on time for 9.9 years...
In my experience, it's better to avoid using two SSDs from the same vendor or series. If it goes badly, both SSDs die at the same time and your raid is gone.
If you dont have any other option, I highly suggest monitoring the box closely (both mdadm / smart values) and implement daily or at a minimum weekly backups. Thats all what you can do with a rented box.
Problem with drive failures is they are completely unpredictable.
IMO the question about raid or no raid is the question about availability you need. Can you afford to wait for drive replacement and backup recovery with things offline? If yes you can use single drive with proper backups safely. If no - raid will help in many cases, but another methods like having a second server may be even better.
Also, in terms of "likeliness" of failure - drives fail, sometimes because of manufacturing flaws, sometimes because of wear, sometimes for no identifiable/predictable reason (and with no prior indication). It is more of a question of "when" than "if".
This is like asking whether you are going to have a heart attack soon.
They provided me with this SSD about two months ago, I’m kinda curious what the last guy did to the thing. But did get almost new hard drives on this auction box.
what number is high here that gives you concern? Is it Total Host Writes?
2PB writes, 126 reallocated and 5% wear leveling count remaining, that disk is pretty much dead.
Try dropping it. I had a ~2 year spin time 2.5 completely die when detached (in padded external case) and falling about 60cm onto carpet. Only took once, so I guess I won the lottery.