New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Retrieving speedtest.net configuration...
Retrieving speedtest.net server list...
Testing from Tiscali France (212.47.239.222)...
Selecting best server based on latency...
Hosted by FreeMobile (Paris) [1.59 km]: 5.087 ms
Testing download speed........................................
Download: 336.16 Mbits/s
Testing upload speed..................................................
Upload: 137.96 Mbits/s
Windows Server RT available?
no centos?
I smell a great seedbox opportunity. At least when online.net makes this public.
Edit: As a bored student who should be studying,
Apparently running KDE works fine. http://imgur.com/L4meJNC
Load is a bit high (at 1.x) even after nepomuk has finished scanning
Further Edit:
Just noticed that there is a HDD as well as ndb1
Nah, you can't install third party programs.
These are SAN powered I believe, SAN is accessed over the 1Gb/s port.
So is there any information or even rumors on pricing?
Got myself invited. The CPU load never goes below 1.1. The interface is slick and doesn't lack options, from simple to advanced. Its a fine alternative to OVH OpenVZ line and RunAbove, but for anything critical mission I would stay away. At least for now.
Do have some invites, PM with your name and e-mail address and I'll add you to the queue.
That's probably something being misconfigured in the kernel, I had this on another ARM platform (sunxi, aka Cubieboard and friends), where the kernel was [from what I could gather] waiting on some USB-related lock, which never got released. Did not affect the performance in any way; so very likely something similar and just cosmetic here, too.
I sent an invitation request yesterday... I got it today
anyone managed to get a 'reserved' IP rather than a dynamically assigned one?
starting/allocating node doesn't work well now, maybe they are short on ressources?
panel an configurable options looks promising though, now curious about the pricing too...
@Falzo, do not use a 60GB LSSD, use a LHDD - they are really short on SSD space. It doesn't boots up cause it can not allocated SSD space anymore
thanks for that advice...
just trying to do some lvm configuration with multiple attach storages - not that easy as there seems to be a bug not fixed in debian wheezy package of lvm2, but jessie looks good...
just in case someone is interested in unixbench results of the ARM processor:
Benchmark Run: Sa Nov 15 2014 13:05:27 - 13:34:18 4 CPUs in system; running 1 parallel copy of tests System Benchmarks Index Values BASELINE RESULT INDEX Dhrystone 2 using register variables 116700.0 3815995.4 327.0 Double-Precision Whetstone 55.0 552.6 100.5 Execl Throughput 43.0 497.5 115.7 File Copy 1024 bufsize 2000 maxblocks 3960.0 108740.9 274.6 File Copy 256 bufsize 500 maxblocks 1655.0 30217.5 182.6 File Copy 4096 bufsize 8000 maxblocks 5800.0 238377.9 411.0 Pipe Throughput 12440.0 284853.4 229.0 Pipe-based Context Switching 4000.0 43359.0 108.4 Process Creation 126.0 675.3 53.6 Shell Scripts (1 concurrent) 42.4 1119.2 264.0 Shell Scripts (8 concurrent) 6.0 271.2 452.0 System Call Overhead 15000.0 759195.1 506.1 ======== System Benchmarks Index Score 207.3 ------------------------------------------------------------------------ Benchmark Run: Sa Nov 15 2014 13:34:18 - 14:03:20 4 CPUs in system; running 4 parallel copies of tests System Benchmarks Index Values BASELINE RESULT INDEX Dhrystone 2 using register variables 116700.0 15234299.5 1305.4 Double-Precision Whetstone 55.0 2214.1 402.6 Execl Throughput 43.0 1009.8 234.8 File Copy 1024 bufsize 2000 maxblocks 3960.0 141574.5 357.5 File Copy 256 bufsize 500 maxblocks 1655.0 41403.5 250.2 File Copy 4096 bufsize 8000 maxblocks 5800.0 302546.2 521.6 Pipe Throughput 12440.0 919830.0 739.4 Pipe-based Context Switching 4000.0 154333.6 385.8 Process Creation 126.0 2067.0 164.1 Shell Scripts (1 concurrent) 42.4 2137.6 504.2 Shell Scripts (8 concurrent) 6.0 280.7 467.8 System Call Overhead 15000.0 2051978.4 1368.0 ======== System Benchmarks Index Score 458.8benched on runabove sandbox instance too, just for comparison:
Benchmark Run: Sa Nov 15 2014 13:08:26 - 13:37:08 1 CPU in system; running 1 parallel copy of tests System Benchmarks Index Values BASELINE RESULT INDEX Dhrystone 2 using register variables 116700.0 40234485.8 3447.7 Double-Precision Whetstone 55.0 3557.6 646.8 Execl Throughput 43.0 6694.6 1556.9 File Copy 1024 bufsize 2000 maxblocks 3960.0 1233557.7 3115.0 File Copy 256 bufsize 500 maxblocks 1655.0 434201.9 2623.6 File Copy 4096 bufsize 8000 maxblocks 5800.0 3106308.1 5355.7 Pipe Throughput 12440.0 2800174.0 2250.9 Pipe-based Context Switching 4000.0 502641.6 1256.6 Process Creation 126.0 21921.7 1739.8 Shell Scripts (1 concurrent) 42.4 8109.2 1912.6 Shell Scripts (8 concurrent) 6.0 1089.2 1815.4 System Call Overhead 15000.0 5749044.6 3832.7 ======== System Benchmarks Index Score 2153.6This sounds good. Would this be enough to run a mail server and owncloud (for only one person)?
But practically am facing real issues with using them
Should do it quite easily.
Been testing out these boxes in various scenarios as was expecting there to be issues given they are ARM, but not encountered one yet, seems someone (or online themselves as they seem to use their own repositories) have done a lot of work getting these working.
You do realize the 2000 result is not from these ARM servers, right?
this. second benchmark is from runabove labs (OVH) not online.net...
sorry if that caused confusion, edited post and made it bold to make it clearer ;-)
Has anyone noticed that
a) The hostname gets reset daily (Well, I dont know if it is daily, but it gets reset the next day when I wake up)
b) The nameservers get reset daily as well
You can fix both by removing a script file which populates them, think there is a hot link on the control panel above hostname explaining how to fix the hostname one.
God knows why they do it, guess its one thing they will remove before it goes 'live'.
Should say really love these little boxes, and got a feeling they are going to price them VERY cheaply, as worked out each one costs under $50 to build once you factor in they have built 50,000+ of them. I wouldn't be shocked at an opening price for these first gen boxes of just a 5 euro a month just to get a market and traction behind the idea, with later ones with a more modern ARM CPU being higher.
Hmm, I cant find the script file though I might be looking in the wrong place
How do these stack up against the performance of their servers with VIA processors?
"Oops! All C1 servers are busy, please retry in a few minutes
"
Took a bit of tinkering, but I finally was told (and found) that the script was in the exit hooks folder in /etc/dhcp/dhclient*
I removed it, and rebooted, and the hostname was still there. Tinkered a bit more around with dhclient, and I removed the dns and domain name from the configuration, and the DNS servers did not change as well.
Looks like these are getting closer to release.
Just had an email saying all accounts have to have a valid phone number attached to them, and you have 72 hours to add it. The date when these amazing little boxes go live comes closer, and assuming the price tag is as low as I think it will be I will be definitely be using them.
@frank and others
Why do you think that these will be so damn cheap? I think they never talked about the price?
It was said in a news article that they'll be competing with DigitalOcean. It was also said that they were looking to release mid-January.
I don't think these can be more expensive than their own SCGen2 (6 EUR/mo).
Each individual core of the four is much slower than the Via Nano.
And they said multiple times that the pricing will be "comparable to DigitalOcean", which many people automatically equated with DO's $5 plan.
5 EUR is still too much, 2-3 EUR sounds more like it.
Because these were designed for quick, easy, dedicated deployment by online.net. We're talking about a place where an 8 core box with 8GB of memory is 15 euro. These being anything but cheap would be the more strange proposal given the vendor's existing pricing structure.