Howdy, Stranger!

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


i-83.net NAT storage VPS performance?
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.

i-83.net NAT storage VPS performance?

Does anyone here have stats/info to share regarding the disk I/O performance of i-83.net's NAT storage VPS range? I'm looking at one of these as a combination backup/torrent server so it will need decently fast random and sequential disk I/O.

Comments

  • emmd19emmd19 Member
    edited January 2017

    Thanks. Would you happen to have random R/W benchmarks, e.g. sysbench? https://wiki.mikejung.biz/Sysbench#VPS_and_Cloud_Server_Random_Read_Performance_Tests_Using_Sysbench

  • user123user123 Member
    edited January 2017

    @emmd19 I have one. I'm hoping to buy more if they have another sale like their one last month (they were OOS of the location I wanted and the price had gone up by the time I could get one that had been restocked). Network-wise, it's been quite good. However, the server has been laggy more than expected while I've been logged in/interacting with it. That is, there is sometimes a large lag between typing a command into my terminal and it actually appearing on the screen or being carried out, even when my line is unused. Sometimes, it affects download speeds. I've noticed that when I was transferring files from other servers, speeds would be very variable (from 500kB/s to 8MB/s) and sometimes even go to zero. My guess is that it was because the disk was overwhelmed (there would be a lot of lag at that time), but that's just my guess. Honestly, they're so cheap I would buy more during the next sale because I need the extra storage for overflow files.

    I don't think they'd be keen on you torrenting on it, unless it was just Linux ISOs and you drastically throttled the reads, writes, connections, and speed.

  • i83i83 Member

    emmd19 said: Would you happen to have sequential R/W benchmarks, e.g. sysbench?

    Not to hand, If you let me know which location/package you are interested in I can ask one of the tech's to run the benchmark for you shortly.

  • @i83 said:

    emmd19 said: Would you happen to have sequential R/W benchmarks, e.g. sysbench?

    Not to hand, If you let me know which location/package you are interested in I can ask one of the tech's to run the benchmark for you shortly.

    Thanks! Also that should read random, not sequential. I'm interested in the S250 package @ CA/BHS.

  • @i83 any plans on adding more NAT locations such Romania, Russia, Hong Kong, or some other exotic locations? I have 5 NATs with you guys and want to add more but in other locations

  • i83i83 Member

    user123 said: I have one.

    Open a ticket if its still an issue and we can take a deeper look.

    emmd19 said: Thanks!

    No problem, added to internal tasklist for when one of the techs has a moment.

    Cyph3r said: any plans on adding more NAT locations

    Maybe, we will have to see what 2017 holds...

  • Cyph3rCyph3r Member
    edited January 2017
  • @i83 said:

    user123 said: I have one.

    Open a ticket if its still an issue and we can take a deeper look.

    Thanks, will do. It is intermittent and I had avoided opening one because it's a NAT plan (force of habit from LES), but I will open a ticket if it keeps happening. The bench (below) I ran after typing this out does point to the disk I/O potentially being a problem.

    @emmd19 bench from their NC location as of just now:

    root@vps:~# wget freevps.us/downloads/bench.sh -O - -o /dev/null|bash
    Benchmark started on Mon Jan  9 17:32:24 EST 2017
    Full benchmark log: /root/bench.log
    
    System Info
    -----------
    Processor       : Intel(R) Xeon(R) CPU           L5420  @ 2.50GHz
    CPU Cores       : 1
    Frequency       : 2493.735 MHz
    Memory          : 512 MB
    Swap            : 512 MB
    Uptime          : 1 day, 1:56,
    
    OS              : Debian GNU/Linux stretch/sid
    Arch            : x86_64 (64 Bit)
    Kernel          : 2.6.32-042stab113.11
    Hostname        : vpsnc
    
    
    Speedtest (IPv4 only)
    ---------------------
    Your public IPv4 is 127.0.0.1
    
    Location                Provider        Speed
    CDN                     Cachefly        38.7MB/s
    
    Atlanta, GA, US         Coloat          8.66MB/s
    Dallas, TX, US          Softlayer       8.51MB/s
    Seattle, WA, US         Softlayer       4.27MB/s
    San Jose, CA, US        Softlayer       5.49MB/s
    Washington, DC, US      Softlayer       15.5MB/s
    
    Tokyo, Japan            Linode          2.97MB/s
    Singapore               Softlayer       2.31MB/s
    
    Rotterdam, Netherlands  id3.net         3.89MB/s
    Haarlem, Netherlands    Leaseweb        4.98MB/s
    
    
    Disk Speed
    ----------
    I/O (1st run)   : 21.5 MB/s
    I/O (2nd run)   : 48.8 MB/s
    I/O (3rd run)   : 48.9 MB/s
    Average I/O     : 39.7333 MB/s
    
  • i83i83 Member

    The bench (below) I ran after typing this out does point to the disk I/O potentially being a problem.

    @user123 Wild guess your on nc7? - It is currently undergoing a RAID rebuild so performance is down on the disks over standard which may explain the lag you mentioned.

    If its still persistent once the rebuild is complete (can be checked via the status page) then open up a ticket and we can investigate further.

  • @i83 Probably (external IP ends in 71), though unlike the lag I was talking about before, I'm not getting any lag right now when entering commands. I'm sure it will be fine.

    Thanked by 1i83
  • i83i83 Member
    edited January 2017

    @emmd19 - S250, CA

    freevps.us

    [root@testSTOR250 ~]# wget freevps.us/downloads/bench.sh -O - -o /dev/null|bash
    Benchmark started on Mon Jan  9 18:28:49 EST 2017
    Full benchmark log: /root/bench.log
    
    System Info
    -----------
    Processor       : Intel(R) Xeon(R) CPU E3-1245 V2 @ 3.40GHz
    CPU Cores       : 1
    Frequency       : 3399.992 MHz
    Memory          : 512 MB
    Swap            :  MB
    Uptime          : 7 min,
    
    OS              : \S
    Arch            : x86_64 (64 Bit)
    Kernel          : 2.6.32-042stab116.2
    Hostname        : testSTOR250
    
    
    Speedtest (IPv4 only)
    ---------------------
    Your public IPv4 is xxx.xx.x.xxx
    
    Location                Provider        Speed
    CDN                     Cachefly        92.8MB/s
    
    Atlanta, GA, US         Coloat          23.4MB/s
    Dallas, TX, US          Softlayer       46.7MB/s
    Seattle, WA, US         Softlayer       21.7MB/s
    San Jose, CA, US        Softlayer       28.5MB/s
    Washington, DC, US      Softlayer       48.4MB/s
    
    Tokyo, Japan            Linode          11.4MB/s
    Singapore               Softlayer       7.59MB/s
    
    Rotterdam, Netherlands  id3.net         8.00MB/s
    Haarlem, Netherlands    Leaseweb        18.9MB/s
    
    
    Disk Speed
    ----------
    I/O (1st run)   : 90.7 MB/s
    I/O (2nd run)   : 73.9 MB/s
    I/O (3rd run)   : 102 MB/s
    Average I/O     : 88.8667 MB/s
    

    Sysbench

    [root@testSTOR250 ~]# sysbench --test=fileio --file-total-size=1G --file-test-mode=rndrw --max-time=120 --max-requests=0 --file-block-size=4K --file-num=64 --num-threads=1 run
    sysbench 0.4.12:  multi-threaded system evaluation benchmark
    
    Running the test with following options:
    Number of threads: 1
    
    Extra file open flags: 0
    64 files, 16Mb each
    1Gb total file size
    Block size 4Kb
    Number of random requests for random IO: 0
    Read/Write ratio for combined random IO test: 1.50
    Periodic FSYNC enabled, calling fsync() each 100 requests.
    Calling fsync() at the end of test, Enabled.
    Using synchronous I/O mode
    Doing random r/w test
    Threads started!
    Time limit exceeded, exiting...
    Done.
    
    Operations performed:  16800 Read, 11200 Write, 17878 Other = 45878 Total
    Read 65.625Mb  Written 43.75Mb  Total transferred 109.38Mb  (933.33Kb/sec)
      233.33 Requests/sec executed
    
    Test execution summary:
        total time:                          120.0010s
        total number of events:              28000
        total time taken by event execution: 0.0909
        per-request statistics:
             min:                                  0.00ms
             avg:                                  0.00ms
             max:                                  0.29ms
             approx.  95 percentile:               0.01ms
    
    Threads fairness:
        events (avg/stddev):           28000.0000/0.00
        execution time (avg/stddev):   0.0909/0.00
    
  • @i83 Thank you. To be honest, that is a bit slower than I was hoping. Would you say that is the normal level of performance I can expect out of it? Would any of the higher-tier storage options be faster? I ask because my current 200GB KVM and RAID10-based storage VPS from HostUS (which only costs a little bit more) benched the following using the same parameters:

    sysbench 0.4.12:  multi-threaded system evaluation benchmark
    
    Running the test with following options:
    Number of threads: 1
    
    Extra file open flags: 0
    64 files, 16Mb each
    1Gb total file size
    Block size 4Kb
    Number of random requests for random IO: 0
    Read/Write ratio for combined random IO test: 1.50
    Periodic FSYNC enabled, calling fsync() each 100 requests.
    Calling fsync() at the end of test, Enabled.
    Using synchronous I/O mode
    Doing random r/w test
    Threads started!
    Time limit exceeded, exiting...
    Done.
    
    Operations performed:  375540 Read, 250360 Write, 400537 Other = 1026437 Total
    Read 1.4326Gb  Written 977.97Mb  Total transferred 2.3876Gb  (20.368Mb/sec)
     5214.28 Requests/sec executed
    
    Test execution summary:
        total time:                          120.0358s
        total number of events:              625900
        total time taken by event execution: 41.2986
        per-request statistics:
             min:                                  0.00ms
             avg:                                  0.07ms
             max:                                714.09ms
             approx.  95 percentile:               0.09ms
    
    Threads fairness:
        events (avg/stddev):           625900.0000/0.00
        execution time (avg/stddev):   41.2986/0.00
    
  • i83i83 Member
    edited January 2017

    emmd19 said: To be honest, that is a bit slower than I was hoping. Would you say that is the normal level of performance I can expect out of it?

    Little slower than usual but the node is quite busy today. That is also our smallest storage VM which is assigned the lowest I/O priority.

    Would any of the higher-tier storage options be faster?

    Yes, also the OVH nodes are generally slower on I/O over our other locations. Alternatively you could look into the Premium KVM range from QuadHost rather than a NAT solution.

    I ask because my current 200GB KVM and RAID10-based storage VPS from HostUS
    benched the following using the same parameters:

    Depending on the RAM allocation (as per the link below) you may also still be in cache, as per your linked document the value was set double the RAM offered on our service.

    (which only costs a little bit more)

    Forgive me but their site lists a 200G KVM as $70/m? https://hostus.us/kvm-vps.html .

    Our product is £20/yr (~$24).

  • emmd19emmd19 Member
    edited January 2017

    Ah, gotcha. I forgot to mention that the HostUS plan I'm on was a 256MB RAM LEB special, which was pretty cheap with annual billing: https://lowendbox.com/blog/hostus-new-storage-kvm-line-starting-10quarter-asia-openvz-25year-8-locations/

    Can you give me an idea of how the higher-tier plans perform? My second choice would be the S250 or S500 plan in London.

    As for the Premium KVM line, I see you a have a custom option - would it be possible to provision something similarly, i.e. 256-512MB of RAM and ~200GB spinning disk? How much would that run me monthly or annually?

    Sorry for all the questions, I'm just trying to do some research before I buy :v

  • BeardyUnixGuyBeardyUnixGuy Member
    edited January 2017

    I had a jolly old chuckle when I read "Victim of its Success" on @i83's website. Thanks @i83!

    Thanked by 2Clouvider i83
  • williewillie Member
    edited January 2017

    I have 2 of these servers and could you guys please take it easy on them? They're storage servers, I like to think not intended for running databases or torrents. I like them a lot but the idea is to park semi-cold data on them and get it again later. The iops available is not all that high and doesn't need to be. I do sshfs mount them for convenient serial access but don't expect that to be as fast as local files.

    The usual signs are that if you type "ls" or the like, there's usually a little pause before output, that you rarely experience with SSD servers. If they beefed up the servers to get rid of that pause, it would cost more without doing its intended function any better, and I'd rather not pay for that.

    If you want cheap network SSD storage, try Scaleway, 2 euro/month per 50GB chunk. Not like a local NVMe but still faster than any HDD. I think they advertise 2000 iops per chunk.

    Thanked by 1PieHasBeenEaten
Sign In or Register to comment.