I had installed that file server on an M.2 memory stick (500 gig) on the motherboard and plugin itself, so I never used a PCI card adaptor.
How is that possible? I thought this motherboard has Key E M.2 and there are no SSDs in that format?
Anu speculations when the next generation of RK3399 chip will be released? Or it will be the best one for a long time?
My goal is also to use my existing 2.5" disks and not to lose data if one or two of them fail. I'm sure the result still can be relatively compact, quiet and low power, especialy since most of the time it will be idling.
Nefertiti, your first links leads to a 404 page…
But your setup seems rather simple and elegant. Is there any other ARM board that can take PCI cards?
If I foresee getting two more 2.5 disks in the future, I just get a PCI SATA card with more connectors, right?
Are there any drawbacks for the setup Nefertiti uses?
Not really since the count of disks is the problem to get something 'compact, quiet and low power' and inexpensive as well.
Do you mean that there's no ARM board that can be easily / natively connected to 4 SATA disks?
I have four 2,5" SATA disks (500GB-1TB) and would like to make a small NAS out of them. What kind of hardware would you suggest for that? I would only use it to store my files, no other functions needed. I can do withouth Wifi, Ethernet is enough for me. Speed is unimportant, I just need my NAS to be compact, quiet and low power.
I guess ARM would be more suitable for such task than x86, right?
Has anyone found cheap Denverton motherboards? The processors itself are quite inexpensive (starting from $27) so I thought they could be useful for a lowend NAS, but no many motherboards with them so far.
Found only http://www.supermicro.com/products/motherboard/ATOM/ but these are expensive, starting from 250 EUR. And I don't need 4 x 1GbE anyway, however 8 x SATA3 would be useful.
Is it because these chips were released only a month ago? Or are they targeted only for high end? I don't know much about Intel product lines.
Thanks, MergerFS + SnapRAID might fit what I'm looking for!
Can they be used together with OMV? Or maybe I don't even need OMV if I put MergerFS + SnapRAID on stock Debian and turn on some kind of file sharing?
I have a backup server (an R-PI with a 4TB usb drive) that duplicates my 24x7 main server once a week
Fine if your backup fits on one disk.
I don't have a NAS yet. I'm planning to build one and appear as ~15 GB single volume on my computer when connected. Can't do that without RAID or ZFS/BTRFS to merge several HD into one logical one.
Looks like ZFS has at least as much oportunities to fail as RAID if not more. This is what I read on FreeNAS forum about ZFS:
For maximum performance and reliability, you should never try to use ZFS with less than 8GB of RAM and a 64-bit system under any circumstances, regardless of the size of your pool.
We typically see at least 1-2 people every week that break this rule and they lose their pools suddenly. There is no recovery if you damage your pool and ignoring this warning.
NAS4Free recommends 1GB per TB of storage without the baseline 8GB. So if you have 5TB of storage you’d need 13GB of RAM for FreeNAS but only 5GB for NAS4Free.
ZFS has very few “recovery tools” unlike many other file systems. For this reason, backups are very important. If the zpool becomes unmountable and cannot be repaired there are no easy software tools or reasonably priced recovery specialists you can use to recover your data.
Setup the UPS service (and use an UPS). Many people have corrupted their zpool because of an improper shutdown.
You cannot add more hard drives to a VDev once it is created.
Once you add a VDev to a zpool it cannot be removed for any reason. Even if you “just” added it because of a typo at the command line or a mistake in the UI.
Failure of a VDev in a zpool will cause the zpool to fail. This means all data in the zpool is lost. There is no chance of recovering data from the remaining VDevs and there are no recovery tools for ZFS.
So ZFS seems very prone to failure if you don't spend thousands of $$$ on hardware, tens of GB of ECC RAM etc. And very inflexible.
Is there a system designed where for example if I choose 25% redundancy and have 4 HDs, if any one fails, I lose no data? If 2 fail, I still can recover 66% of my data. If 3 fail, I can recover 33% of the data. Neither RAID nor ZFS/BTRFS can provide that to my knowledge, the fails result in catastrophic loss of all data. It shouldn't be like this, should it?
It appears that there is no solution that would allow me in case of failure plug the still undamaged disks into my computer one by one and recover what is still recoverable, be it RAID, ZFS or BTRFS.
So you actually don't know what happened?
Can someone explain what exactly went wrong in this case and how can others avoid that?
To put it more precisely — are there any plans to use Debian 9 for OMV 3 in the future? From the replies here looks like OMV 3 will remain on Debian 8 forever, right?
Now that Debian 9 Stretch is released as stable, are there any plans to use it instead of Debian 8 for OMV?
I also have a question about this motherboard.
It has a Mini PCI-Express slot. Can I put a SSD disk in it? It would be very convenient not to waste a SATA port.
I'm trying to build very-low-budget NAS (purely for backup / storage, no transcoding or anything like that). Seems that Q1900-ITX is the best bang for the buck available at the moment? Reasonable performance, CPU included, 70 EUR/USD, 4 SATA ports, low power. (Have some SO-DIMMs laying around as well, nice that I can reuse them.)
Is there any motherboard with much better performance, more SATA slots or lower power available for <100 EUR/USD? (Or maybe not now but in near future.) From my research it seems there is none, or have I missed any?
Thanks in advance!