DIY NAS. RPi4B?

Jun 30, 2019 Last reply: 7 years ago 44 Replies

Quite. I had to work quite hard to build my server and keep the power consumption down.

I did it by carefully choosing an Atom powered ITX board, using 3 x

500GB laptop drives and making it passively cooled (every fan adds quite a bit).

It also hibernates as soon as the last user logs off (and any torrent file finishes or the network access drops below a minimum usage or there is no guest access etc) and the hdd's idle when not being used etc.

So, whilst not quite as low as say a DS218j, it's much closer than one of TNP's 'old PC's' is likely to be ... and every little helps (lower the background level). ;-)

When measuring desktop PC's in general I was horrified just how much power some of them consumed, even when idle! ;-(

Daughter was saying last night that she wishes she had all her stuff in one place again (as she used to when she lived here) as she is often on her phone, laptop or desktop. I would love to give her something like a 218j but we can't really afford it atm (funeral / other costs etc). So I could build her a RPi based unit and it probably wouldn't need to be massive, so even 1TB (or whatever the deal / sweet spot was for USB3 portable drives) of storage would probably do, but it really would need to have 1) some sort of RAID(1) and 2) the content of the drives would need to be readable in the event of a system hardware failure (on another box etc).

I'm currently testing a new 32GB uSD card and will install Rasbian on it. I would then like to install the latest version of OMV and (try to) configure it for RAID1 (I don't think there are any OMV images for RPi4 yet).

Performance doesn't really matter (within reason) and as long as the soft RAID doesn't impact the performance too much, it shouldn't be any worse (and likely to be much better) than my OMV running on a RPI2B. ;-)

OR, if RAID isn't advised, even on a RPi4, working out how you could automate the backup of a single shared volume onto an alternative one? Whilst this wouldn't be a production solution, availability would be important as stuff generally goes wrong when time is critical or the data most important. ;-(

Cheers, T i m

Bullshit. None of mine have ever failed and some are much older than that. In continuous use too.

Yep, only ever had the one fail.

In my working life I have run an ISP with dozens of servers.

I have a close relationship with a PC builder as a result. And a good sttaistical knowledg of disk lives

Disks that are thrashed 24x7 will barely make a year warranty. Normal disks last 5-7.

Disks that are seldom used can go on for 10 or more.

I know what I am talking about.

In my server after about 14 years I have replaced 3 disks

I have had three fail. At home.

Burned out a CPU or two in the business fromn dead fans.

Mostly something like ethernet stops working, or onboard video or sound goes.

Only if they arent properly cooled. And that magnificent statistical report from that massive great server farm shows nothing even remotely like that claim of yours.

formatting link

Doesn?t show that either.

You clearly don?t.

That?s nothing like your original claim.

Then you must be using crap motherboards.

Never had a dead cpu fan either, but then I use intel cpus.

You must be using crap motherboards.

I originally had 1U servers, all fitted with ITX boards. They worked fine, but the fans soon became noisy and they had problems in the height of summer.

I ended up replacing them with HP microservers. A bit more power consumption, but better for my hearing.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required