What kind of SSD for Raspberry Pi when using persistence? Sata, mSata or m.2?

Hi,
i wanted to use persistence with influxDB and grafana. I did read that the microsd cards break fast. So I am wondering if I should copy the filesystem to a SSD.

Unfortunately I am not sure what kind of SSD I should take. Sata, mSata or m.2?

Is mSata and m.2 in a usb case even compatible to the Pi?

Is it a good idea to use one of these usb-sticklike m.2 cases and stick them in the Pi, or are they too big and will block the usb ports?

I think I would prefer mSata.They have nice shape and are small.
Could I just take a random chinese mSata? I have 32gb sd card. Can use a 16gb mSata with openhabian-config, or do I need a workaround?

va

I just use one with a usb to 2/5" sata adapter with a 275gb Crucial SSD

If you’re using m.2/mSata, how are you physically connecting it to the Raspberry pi?

And if it’s in any way attached to the usb bus, you’re restricted there for performance, but as you say, you’re not really doing it for performance, but for longevity.

20 bucks for a 32gb mSata: e.g. Kingspec 32GB mSata

and then 10 bucks for a case like that
41Ywhz401dL

then connect it with the cable to the usb port.

Is a bigger size than 32GB mabye better concerning longevity?

Does anybody know if this configuration will work at all?

Yes, the concern is the quality of the drive and the drive controller.
The drive controller is everything when it comes to SSD longevity. It managed the flash cells to “rotate” data is order to minimise wear.
I am aware that price is an issue it but look for a known brand

The RPI has to connect via USB. There is no SATA controller.

1 Like

The type of SSD doesn’t really matter much. You can even use NFS.
I’d be more afraid of running OH and Grafana on the same RPi. I’d use a second one.
And be aware that there’s currently a number of problems with Grafana/PhantomJS. Search the forum for more info.

1 Like

Yes it works:

The speed doesn’t matter much. You are limited by the speed of USB bus, which is slower than any drive. So a m.2 drive will not give any better performance, and using a m.2 drive to achieve better speed performance won’t work.

1 Like

Does graphana consume so much resources?
I just want to display temperature and humidity of three sensors. Somebody on the board told me to use influxdb and graphana.
What is a good alternative to graphana and a second Rpi?

You can use the built it graphing capabilities on openHAB, influx and grafana are nice to have but not necessary for graphing, especially when your scope is limited like yours

1 Like

I accidentially answered the wrong thread, but long story short:

  • use big HD volumes (to decrease Change of overwriting same sectors over and over again)
  • use old-fashioned HDs instead of cheap SSDs
  • use the built-in “boot from USB”-Option of the newer Pis

…and here’s the Long answer in the wrong thread:

2 Likes

I am not sure if I understand completely, so you choose old-fashioned HD because it is more reliable than SSD or because it is cheaper?

Are you using persistence?

Did you separate the logs and swapping files like mentioned in this post: Corrupt FileSystems every 2-3 month?

Or is this separation not needed anymore with an old fashioned HD?

Do you know what is the power consumption difference between a SSD and the old fashioned HD? Didn’t find anything conclusive…

So I hope! :wink: (reliability)

yes - MapDB and for some items RRD. But the main persistence is a MariaDB on my NAS.

nope. I do backup regularly (only config, still trying to figure Amanda out!)=

I don’t know… :wink:

hmm… I’m not sure, I think I read sometime ago, that for spinup the HD needs a bit more power, but in “real life” it wouldn’t make much difference?
My HD-housing came with an Y-cable so I have a seperate USB-Port just for power (I could also use some PIN with 5V power. But I already use a powered USB-Hub which works fine.

hope that helps…

1 Like

I’m using swissbit industrial microSD Cards wich have power-fail-safe capabilities.
I had no problem with these the last 2 years.

1 Like