NAS appliance or home server
in [UK Linux]
|
Prev: Linux Laptops
Next: I've finally snapped
From: jasee on 29 Aug 2008 08:33 Gordon Henderson wrote: > In article <EvCdnSn1ttbgnyrVnZ2dnUVZ8qXinZ2d(a)bt.com>, > jasee <jasee(a)btinternet.com> wrote: >> Gordon Henderson wrote: >>> In article <313030303737303648B5E26D66(a)plugzetnet.co.uk>, >>> Johnny B Good <jcs.computersbutt(a)plugzetnet.co.uk> wrote: >>> >>>> I must admit, a room thermometer is not particularly high on most >>>> folks "Must Have" list (myself included :-), but unless you have >>>> the means to monitor room temperature (ambient, as far as >>>> electronic kit is concerned), all those temperature readouts on >>>> the critical parts of a PC have their value as a diagnostic >>>> indicator compromised. >>> >>> At bit OT, but I got myself one of those IR spot thermometers a >>> while back when I started out building my own PBX/NAS/Router boxes >>> just to make sure that inside their fanless environment nothing was >>> getting too hot - since then I've used it on all sort of things... >>> The one I have is: (Hm. not in-sock anymore, but similar to) >>> >>> >> http://www.maplin.co.uk/Module.aspx?ModuleNo=223009&criteria=thermometer&doy=28m8 >> >> Interesting but surely these things measure the temperature on the >> outside surface of the box which is probably a lot cooler than >> inside? > > Yes, they do, but I was running with the lids on, then removing it > and measuring immediately to try to get worst-case values. Maybe not > that scientific, but it seemed to satisfy me :) > > I'm currently frowning over a new box with 2 x 1TB WDC "Green" drives > which are running at 50C surface temp (52 according to smartctl) which > is annoying as I think I'm going to have to re-position them. > > Maybe I'd have been better off without it! Wouldn't it be possible to simply add a case fan or two inside the box.
From: Gordon Henderson on 29 Aug 2008 09:32 In article <qNCdnUoWRqUZcSrVnZ2dnUVZ8g6dnZ2d(a)bt.com>, jasee <jasee(a)btinternet.com> wrote: >Gordon Henderson wrote: >> In article <EvCdnSn1ttbgnyrVnZ2dnUVZ8qXinZ2d(a)bt.com>, >> jasee <jasee(a)btinternet.com> wrote: >>> Gordon Henderson wrote: >>>> In article <313030303737303648B5E26D66(a)plugzetnet.co.uk>, >>>> Johnny B Good <jcs.computersbutt(a)plugzetnet.co.uk> wrote: >>>> >>>>> I must admit, a room thermometer is not particularly high on most >>>>> folks "Must Have" list (myself included :-), but unless you have >>>>> the means to monitor room temperature (ambient, as far as >>>>> electronic kit is concerned), all those temperature readouts on >>>>> the critical parts of a PC have their value as a diagnostic >>>>> indicator compromised. >>>> >>>> At bit OT, but I got myself one of those IR spot thermometers a >>>> while back when I started out building my own PBX/NAS/Router boxes >>>> just to make sure that inside their fanless environment nothing was >>>> getting too hot - since then I've used it on all sort of things... >>>> The one I have is: (Hm. not in-sock anymore, but similar to) >>>> >>>> >>> >http://www.maplin.co.uk/Module.aspx?ModuleNo=223009&criteria=thermometer&doy=28m8 >>> >>> Interesting but surely these things measure the temperature on the >>> outside surface of the box which is probably a lot cooler than >>> inside? >> >> Yes, they do, but I was running with the lids on, then removing it >> and measuring immediately to try to get worst-case values. Maybe not >> that scientific, but it seemed to satisfy me :) >> >> I'm currently frowning over a new box with 2 x 1TB WDC "Green" drives >> which are running at 50C surface temp (52 according to smartctl) which >> is annoying as I think I'm going to have to re-position them. >> >> Maybe I'd have been better off without it! > >Wouldn't it be possible to simply add a case fan or two inside the box. Oh sure - but that's not the point. I want (have) a fanless unit (which I use for PBXs) that can go into a small office and be totally silent - so I bought the IR thermometer to check them (and I have another thermometer with an "outdoor" sensor on a bit of wire I use too when I want to keep the lid on a box) It matters not for my own server - it's got a PSU fan anyway. (which is surprisingly noisy ): I've since re-arranged the drives and they're now running cool enough for me to not be concerned about them. # hddtemp /dev/sda /dev/sdb /dev/sda: WDC WD10EACS-00ZJB0: 36 C /dev/sdb: WDC WD10EACS-00ZJB0: 38 C Once upon a time I did get "stung" with a case fan failling - and the server was 400 miles away too, so since then I've a bit fussy about cooling and so on - and if I can get away without fans then I do... Gordon
From: Jonathan Buzzard on 30 Aug 2008 18:51 On Sun, 17 Aug 2008 00:34:29 +0100, Daniel James wrote: > In article news:<2objn5-4qq.ln1(a)small.buzzard.me.uk>, Jonathan Buzzard > wrote: >> My mini-ITX box draw 30W at the wall ... > > Useful, thanks. > > I've been eyeing the EN1200 ... but that is one of the more expensive > mini-ITX boards. What are you using for a case? (just out of interest) The Pack-Box-A4 http://www.icp-epia.co.uk/index.php?act=viewProd&productId=321 With the nasty 40mm fan removed. Need some metal work to use, because the mounting holes for two 2.5" drives precluded actually connecting the drives up... > That's better than any other general purpose system I've looked at, but > noticeably worse than some NAS appliances. > > I've just been reading an article on the Netgear/Irfrant ReadyNAS DUO > which quotes around 27W at idle with two 500GB drives, and 12W with the > drives spun down. Don't believe that. They would appear to be claiming a mear 15W for two 3.5" drives spinning over spun down. I suggest you look at some of the spec sheets for 3.5" drives as that is simply not possible. Two 3.5" drives will take most of that 27W all to themselves when active. By the time you add in losses from the PSU they will take it all and more. >> With the VIA Padlock crypto acceleration, I get some exceptional >> performance out of a 1.2GHz fanless box. > > What does the incarnation of Padlock on that board do? They version I > have (on an M10000) has no acceleration for modular exponentiation, > which you would need for RSA or Diffie-Helmann but they were making > noises about adding a Montgomery multiplier circuit, or something > similar. Has that now happened (I've not been monitoring that)? Makes most (if not all) my crypto stuff go *much* faster. So that is SSL with Apache, SSH and VPN stuff I use. > Is this all working under linux? > Yep. JAB. -- Jonathan A. Buzzard Email: jonathan (at) buzzard.me.uk St. Andrews, United Kingdom.
From: Daniel James on 31 Aug 2008 13:07 In article news:<1gvoo5-a45.ln1(a)small.buzzard.me.uk>, Jonathan Buzzard wrote: > [I wrote:] > > What are you using for a case? (just out of interest) > > The Pack-Box-A4 > > http://www.icp-epia.co.uk/index.php?act=viewProd&productId=321 Ah, interesting. I was aware of the case but for a couple of pounds more from linitx. Not much help if you don't want to use 2.5" drives, though. > > I've just been reading an article on the Netgear/Irfrant ReadyNAS > > DUO which quotes around 27W at idle with two 500GB drives, and 12W > > with the drives spun down. > > Don't believe that. They would appear to be claiming a mear 15W for two > 3.5" drives spinning over spun down. I suggest you look at some of the > spec sheets for 3.5" drives as that is simply not possible. Two 3.5" > drives will take most of that 27W all to themselves when active. Most bog-standard 3.5 drives seem to be spec'd at around 7W when spinning but not read/writing, so I'd say that 15W figure was very reasonable. There are some 'Green' drives available nowadays that use less (the WD drives IIRC quote 4W). > By the time you add in losses from the PSU they will take it all and > more. Yes, the PSU losses will make a difference ... but isn't that always the way? > Makes most (if not all) my crypto stuff go *much* faster. So that is > SSL ... Yes, but, SSL uses asymmetric cryptography for key setup and then uses symmetric cryptography for the actual message encryption. I know the VIA chips provide dramatic speed improvements for symmetric crypto (if you're using AES, anyway) but not whether they yet accelerate the *asymmetric* part. That's what I was asking about. Cheers, Daniel.
From: Daniel James on 4 Sep 2008 09:59
In article news:<1gvoo5-a45.ln1(a)small.buzzard.me.uk>, Jonathan Buzzard wrote: > Don't believe that. They would appear to be claiming a mear 15W for > two 3.5" drives spinning over spun down. I suggest you look at some > of the spec sheets for 3.5" drives as that is simply not possible. > Two 3.5" drives will take most of that 27W all to themselves when > active. By the time you add in losses from the PSU they will take > it all and more. I've been running the ReadyNAS through one of those mains power meters ... not the worlds finest precision instrument, but it seems to give fairly accurate values for known loads. The maximum draw (while booting, spinning up one drive at a time) was 46.6W, the idle current between 22 and 23W, and with the disks spun down around 12W (measured low 11.6W). Pretty-much what Netgear claim. Standby current (when first plugged in, before switching on) was 0.5W. Cheers, Daniel. |