Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Online Lab cloud first hand review
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.

Online Lab cloud first hand review

cnbeiningcnbeining Member
edited March 2015 in Reviews

I just got the invitation from Online.net.

The server I am testing C1, "a 4-cores ARMv7 CPU with 2GB of RAM and a 1 Gbit/s network card."

https://doc.cloud.online.net/faq/server_faq.html

So, basicly this is a real server, not a VPS.

Storages can be in SSD or HDD, but attached via internal network.

Pros:

  • Good peering with EU

  • Scaling

  • No CPU&ram oversell

  • DDOS protected

  • Reserved IP addresses

  • Object storage

  • Quick deploy: as quick as Digitalocean

  • Several one-click install Apps

Cons:

  • Shi**y peering with Asia, even with North America

  • No email sending: get yourself a Mailgun or Mandrill instead(maybe only during the preview?)

  • It's ARM, not x86

Benckmark:

root@labs-d5068e:~# wget freevps.us/downloads/bench.sh -O - -o /dev/null|bash

CPU model : ARMv7 Processor rev 2 (v7l)

Number of cores : 4

CPU frequency : MHz

Total amount of ram : 2020 MB

Total amount of swap : 0 MB

System uptime : 20 min,

Download speed from CacheFly: 57.7MB/s

Download speed from Coloat, Atlanta GA: 4.32MB/s

Download speed from Softlayer, Dallas, TX: 5.88MB/s

Download speed from Linode, Tokyo, JP: 1.56MB/s

Download speed from i3d.net, Rotterdam, NL: 25.8MB/s

Download speed from Leaseweb, Haarlem, NL: 14.2MB/s

Download speed from Softlayer, Singapore: 2.90MB/s

Download speed from Softlayer, Seattle, WA: 323KB/s

Download speed from Softlayer, San Jose, CA: 4.45MB/s

Download speed from Softlayer, Washington, DC: 7.56MB/s

I/O speed : 40.0 MB/s

I will update the serverbear benchmark as soon as it completed.

Update:ServerBear:

http://serverbear.com/benchmark/2015/03/17/7p1bCAzRvTL07W9o

OMG the Unixbench is low...

«1

Comments

  • LeeLee Veteran

    There have been quite a few in depth discussion on this over the past few months.

    Thanked by 1netomx
  • rm_rm_ IPv6 Advocate, Veteran
    edited March 2015

    said: Shi**y peering with Asia

    From DO Singapore to Online.net ARM:

    ...
    HTTP request sent, awaiting response... 200 OK
    Length: 104857600 (100M) [application/octet-stream]
    Saving to: `/dev/null'
    
    100%[=====================================================>] 104,857,600 27.4M/s   in 5.0s    
    
    2015-03-17 20:35:52 (20.0 MB/s) - `/dev/null' saved [104857600/104857600]
  • @rm_ said:

    Quite interesting.

    Benchmarking download from Linode, Tokyo, JP (http://speedtest.tokyo.linode.com/100MB-tokyo.bin)

    Got 147KB/s

    Well....

  • rm_rm_ IPv6 Advocate, Veteran

    cnbeining said: Got 147KB/s

    Getting 1.5 MB/sec from that one, in your 1st post it's 1.56 MB/sec as well.

  • IOPing terrible

    "LOCAL"

    root@arm2:~# ioping / -c 10
    4.0 KiB from / (ext4 /dev/nbd0): request=1 time=1.8 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=2 time=2.3 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=3 time=1.8 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=4 time=1.8 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=5 time=2.4 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=6 time=1.1 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=7 time=1.1 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=8 time=1.1 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=9 time=1.1 ms
    4.0 KiB from / (ext4 /dev/nbd0): request=10 time=1.4 ms
    
    --- / (ext4 /dev/nbd0) ioping statistics ---
    10 requests completed in 9.0 s, 632 iops, 2.5 MiB/s
    min/avg/max/mdev = 1.1 ms / 1.6 ms / 2.4 ms / 457 us
    
    
    DD TEST
    1073741824 bytes (1.1 GB) copied, 14.1266 s, 76.0 MB/s
    

    LHDD

    root@arm2:/# ioping  /mnt/hdd -c 10
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=1 time=1.1 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=2 time=1.1 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=3 time=1.2 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=4 time=1.2 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=5 time=1.1 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=6 time=1.1 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=7 time=1.1 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=8 time=1.2 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=9 time=1.4 ms
    4.0 KiB from /mnt/hdd (ext4 /dev/nbd1): request=10 time=1.1 ms
    
    --- /mnt/hdd (ext4 /dev/nbd1) ioping statistics ---
    10 requests completed in 9.0 s, 858 iops, 3.4 MiB/s
    min/avg/max/mdev = 1.1 ms / 1.2 ms / 1.4 ms / 83 us
    
    
    DD TEST
    1073741824 bytes (1.1 GB) copied, 15.9088 s, 67.5 MB/s
    

    LSSD

    root@arm2:/# ioping /mnt/data -c 10
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=1 time=1.4 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=2 time=1.2 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=3 time=1.1 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=4 time=1.2 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=5 time=1.1 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=6 time=1.1 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=7 time=1.4 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=8 time=1.1 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=9 time=1.1 ms
    4.0 KiB from /mnt/data (ext4 /dev/nbd2): request=10 time=1.1 ms
    
    --- /mnt/data (ext4 /dev/nbd2) ioping statistics ---
    10 requests completed in 9.0 s, 839 iops, 3.3 MiB/s
    min/avg/max/mdev = 1.1 ms / 1.2 ms / 1.4 ms / 117 us
    
    
    DD TEST
    1073741824 bytes (1.1 GB) copied, 14.5826 s, 73.6 MB/s
    
  • DylanDylan Member

    @cnbeining @rm_

    What kernel version do you each have running? Some of the mainline kernels have network (and disk) performance issues.

  • sc754sc754 Member

    How does this arm cpu compare to the RPI 2 I wonder...

  • @Dylan said:
    cnbeining rm_

    What kernel version do you each have running? Some of the mainline kernels have network (and disk) performance issues.

    Default Ubuntu 14.04.

  • DylanDylan Member

    @cnbeining said:
    Default Ubuntu 14.04.

    Type "uname -a" in a console:

    root@c1-10-1-2-16# uname -a
    Linux c1-10-1-2-16 3.19.1-179 #1 SMP Mon Mar 9 12:04:16 UTC 2015 armv7l armv7l armv7l GNU/Linux
  • How many LHDD can be added to a server?unlimited?

  • @Dylan said:
    Linux c1-10-1-2-16 3.19.1-179 #1 SMP Mon Mar 9 12:04:16 UTC 2015 armv7l armv7l armv7l GNU/Linux

    root@labs-d5068e:~# uname -a
    Linux labs-d5068e 3.17.0-119 #1 SMP Thu Nov 20 14:15:44 CET 2014 armv7l armv7l >armv7l GNU/Linux

  • I change to new stable kernel ,

    Linux server01 3.2.34-29 #16 SMP Thu Mar 12 09:45:43 UTC 2015 armv7l armv7l armv7l GNU/Linux

    and I get

  • I received the invite too. Unable to verify phone number. :(

  •  [root@nginx01 ~]# wget http://speedtest.tokyo.linode.com/100MB-tokyo.bin
     --2015-03-18 01:09:52--  http://speedtest.tokyo.linode.com/100MB-tokyo.bin
     Resolving speedtest.tokyo.linode.com... 2400:8900::4b, 106.187.96.148 
     Connecting to speedtest.tokyo.linode.com|2400:8900::4b|:80... connected.
     HTTP request sent, awaiting response... 200 OK
     Length: 104857600 (100M) [application/octet-stream]
     Saving to: “100MB-tokyo.bin”
    
     100%[======================================>] 104,857,600         11.4M/s   in 9.2s
    
  • rm_rm_ IPv6 Advocate, Veteran

    clamhost said: Connecting to speedtest.tokyo.linode.com|2400:8900::4b|

    Are you sure that's on online.net ARM, because those don't have IPv6.

    If not, I wonder what relevance your post has to this thread.

    If yes, you should mention what type of IPv6 tunnel or tunnel broker did you use.

  • dragon1993dragon1993 Member
    edited March 2015

    Nomad said: What a CPU power!

    Weak CPU power.

    C1: ~400

    Kimsufi KS-1: ~600

    DigitalOcean: ~900

  • It's an ARM, not a Ferrari.

    Thanked by 2berkay 5hi74
  • rm_rm_ IPv6 Advocate, Veteran

    dragon1993 said: C1: ~400

    Kimsufi KS-1: ~600

    But it's close to Atom D425 (1-core with HT) http://serverbear.com/1972-ks-4g-yearly-kimsufi

  • @rm_ said:
    But it's close to Atom D425 (1-core with HT) http://serverbear.com/1972-ks-4g-yearly-kimsufi

    Yes, but one powerfull core better than four weak core. I would choose ATOM D425

  • NomadNomad Member

    rds100 said: It's an ARM, not a Ferrari.

    >

    Let me remind, Odroid, Banana Pro etc... There are MANY powerful ARM processors as well. Way more powerful than 400 benchmark on a single core let alone 4 of them.

    This might work if this is released at the price of a Kidechire, mayhaps.
    But other than that it is no good in terms of CPU power. Imagine the performance of apps that run on a single core.

  • DylanDylan Member
    edited March 2015

    @cnbeining said:
    root@labs-d5068e:~# uname -a Linux labs-d5068e 3.17.0-119 #1 SMP Thu Nov 20 14:15:44 CET 2014 armv7l armv7l >armv7l GNU/Linux

    Aha, that's a very old kernel. If you click Show next to Advanced Options in the Online Labs CP for your server there's a Bootscript section -- click Edit there and change the bootscript to either the (stable) or (latest) kernel, then reboot your server and uname -a should show that kernel version. After that, try a speed test again.

  • Dylan said: Aha, that's a very old kernel.

    Slow down young padawan. 2.6.32 is old.

  • jvnadrjvnadr Member
    edited March 2015

    dragon1993 said: Weak CPU power.

    Change kernel.
    When I change it, the benchmark is:

    System Specs
    RAM 2023 MB
    HDD 20 GB
    CPU Model ARMv7 Processor rev 2 (v7l)
    CPU Cores 4
    UnixBench
    UnixBench (w/ all processors) 715.2
    UnixBench (w/ one processor) 262.4

    Much better now! And better than KS

    PS

    I used latest stable 3.19.1-179

  • @jvnadr : How do you change kernel?

  • NeoonNeoon Community Contributor, Veteran
    edited March 2015

    Neat, got my Invite. Lets play with it :D

  • I got one some time ago but the thing that bugs me that some software is still not ARM compatible!

  • NomadNomad Member

    I got two invites...
    Did anyone figure out how to reinstall the os without removing the VPS itself?

    Or should I just create another VPS, install OS, keep the volume and get rid of the VPS instance?

  • @tdttester Do what @Dylan says in his reply:

    Dylan said: Aha, that's a very old kernel. If you click Show next to Advanced Options in the Online Labs CP for your server there's a Bootscript section -- click Edit there and change the bootscript to either the (stable) or (latest) kernel, then reboot your server and uname -a should show that kernel version. After that, try a speed test again.

  • jvnadrjvnadr Member
    edited March 2015

    Nomad said: just create another VPS, install OS, keep the volume and get rid of the VPS instance

    ^^ This (if you created a second volume that can be unmounted).

    TheLonely said: some software is still not ARM compatible

    If you use it for deploy apps for x86 architecture, you cannot use this box. But it is ideal for testing apps that will be used to arm architecture.

    But it is not only for developers.

    I installed a web server and tested there a medium traffic joomla site (~4500 unique visitors per day with flash and javascript parts). It worked like a charm, better that most of vps. Network is very good (at least to EU) and the server behave very well even in peak times (of course, it is very weak for big traffic sites with heavy mysql).

    I also installed nginx rtmp for streaming and used the excellent network behavior of online.net to Greece (in testing traffic, I managed to have 3 simultaneously ~42Mbps uploads to 3 different greek isp's: Otenet, GRnet, Vodafone)

Sign In or Register to comment.