I am torn between utilizing my 250GB M.2 drive or a USB 3.0 thumb drive as my system drive. I am out of locations to add PCIe cards. I only have a 1x spot left and can’t find a solution available to support a M.2 PCIe or PCI-Express 8-Port Internal SATA Pci-E.
I am new to Btrfs and Rockstor and wanted in inquire if if I use a USB thumb drive as my system drive, do I loose much in performance? Are there enough benefits to invest my only M.2 location as a system drive?
Is there is a way to backup my single M.2 drive to a standard HDD if I decide to keep it for storage?
Any help would be greatly appreciated.
rockstor works just fine on a USB3 stick. In fact if it detects that it’s installed on a USB, it will limits the write to the USB.
The next part of your question is not very clear to me. If you have one PCIe 1x slot left then you can definitely find a PCIe 1x to M2 adapter for your card (SATA interface M2).
If you want to make an image of your m2 card then use dd
Make an image of a m2 drive in /dev/hda:
dd if=/dev/hda of=~/m2disk.img
Restore the image to /dev/hdb:
dd if=m2disk.img of=/dev/hdb
@khanh Thank you so much for the reply!
I read that you can use a thumb drive, I was just curious if you invest in an M.2 drive, are there noticeable differences to justify the added cost.
I have not found any 1x to M.2 adapter cards that are non-SATA powered. Everything I am finding is 2x.
Do you know if i do use my M.2 drive for storage, can I back it up to a standard HDD?
Thank you again for the help. I have been waiting to start using Rockstor and want to make sure I set it up right the first time.
you could run it from a floppy drive once it boots up you will not see that much speed improvement (maybe guy) but you should invest money in actual storage speed … or start using ECC ram to keep your data safe.
@Tomasz_Kusmierz Thanks for the help!
Do you backup any M.2 drives? I have a space and wanted to use it for my music collection. I wanted to make sure I can back it up to a standard HDD, as that is my only option.
What do you mean by backing up? and what are you backing up specifically, data or rockstor system files?
I want to back up my MP3 music files on my M.2 drive to another drive. I was not sure if with Btrfs and Rockstor, if I can backup an M.2 drive to an standard HDD drive? Or does Btrfs and Rockstor require an M.2 drive to only be backed up to another M.2 drive?
if you have MP3 files on your M2 drive, you must copy the MP3 files out before using the your M2 drive with Rockstor.
Rockstor will delete all your data the moment you use it with Rockstor. After Rockstor is up and running, you can copy the MP3 files back to your M2 drive (which now becomes part of a NAS).
As far as Rockstor/BTRFS is concern, it doesn’t do “backup”. NAS is not a backup (RAID 1 is not a backup). Having said that you can have 2 different type of HDDs in your Rockstor NAS (SATA + whatever) and use crontab to schedule a data backup from one HDDs to another.
You can look at Rockstor as an App running on top of Centos Linux. In my previous setup, I use MS remote desktop to connect to my Rockstor NAS via xrpd services and everything worked just like a typical Linux workstation.
You should really think on why you want Rockstor. If you want to experiment something new, it’s best to do it on your extra hardware.
[condesending mode on]
So this need to be explained once and for all for everybody who thinks that there is some magical difference between drives.
As far as any modern computational platform is concerned (PC / mobile phone / super computer) everything is just a “address space”.
If you look at what computational platform “thinks with” it’s just a CPU (ALU) with bunch of internal setting that sits somewhere within a contagious address pace with your RAM and anything else that is volatile. As an example a ARM A9 (a 32bit cpu ) can consume a good part of half a GB of address space for just settings (of which you’ve got 4GB due to 32bit only able to address 2^32 = 4GB) and you are left with 3.5GB. Those settings are things that are more or less important … but it’s easier to make CPU able to only access a address space and nothing else and map your settings there.
Now storing stuff is usually separated to different devices, but those again are just “address spaces” … a hard drive will report back cylinders, heads, disks, sectors so the operating system (filesystem driver) is aware of physical layout and is able to optimise access pattern to make it as fast as possible. Still whole disk is just a contiguous address pace starting from ZERO.
When you want to move your operating system from HDD to SDD (considering that latter is always bigger) you can simply copy it byte by byte with:
dd if=/dev/source_dive of=/dev/destination_drive
and everything will just work … you made a perfect clone of it …
And this is regardless of interface you use, M.2, SATA, PATA, SCSI, iSCSI, PXE, ATAPI, FC, SAS, USB, ETH … it is just a space starting from 0 bytes to infinity.
(how you boot from certain device is a different matter but as long as it connects directly to a mother board you can rely within 99% probability that BIOS / UEFI will just work with it)
Now, since you know what a storage device is … file system is a artificial logical entity that is imposing a certain form of organising data. You can’t just take data from one filesystem to another and expect it to work like magic - there has to be some conversion performed there. Conversion tools are very rare, and honestly I would not trust my data with those.
the “condesending mode” is meant as tongue in cheek because there seems to be some marketing ploy that people fall for that aims at making new products seem hyper supa dupa advanced.