Howdy, Stranger!

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

Sign In Register

Quick Links


Categories

In this Discussion

› › Very bad performance downloading from Digital Ocean VPS (HTTP)
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.

Very bad performance downloading from Digital Ocean VPS (HTTP)

krypskryps Member
edited February 2013 in Providers

On my DO "droplet" located in NY I consistently get dismal HTTP download performance. I have tried HTTP download from multiple locations. Here is one test from a BuyVM VPS:

xxx@bvm:~$ wget -O /dev/null http://198.211.100.5/test.dat

... 10,485,760 74.5K/s in 2m 2s

A droplet from the same VM snapshot at the Amsterdam location performs better:

xxx@bvm:~$ wget -O /dev/null http://198.211.124.199/test.dat

... 10,485,760 1.62M/s in 7.6s

Anyone else experiencing similar problems? Do you get the same results when downloading those files.

-- kryps

«1345

Comments

  • SpencerSpencer Member
    edited February 2013

    Delete the droplet then create a new one.

  • krypskryps Member

    @Spencer Did that multiple times even with different droplet sizes. Performance was always similar. Maybe they always stick me on the same bad node.

  • shovenoseshovenose Member, Host Rep

    I am under the belief that DigitalOcean's US bandwidth is extremely oversold. I get horrible transfer speeds. Their Amsterdamn location is probably less popular, and yes it is faster. Scary thing is sometimes it's faster to work on an Amsterdamn node from the US than it is to be on a US node.

  • support123support123 Member

    @kryps I was also facing slow download problem them.I have left them already though I have still 20$ credit to my account

  • krypskryps Member

    @murky Ticket was opened seven hours ago. They told me "We are going to review this." four hours ago.

  • MiguelQMiguelQ Member

    @kryps said: Anyone else experiencing similar problems? Do you get the same results when downloading those files.

    No. From Choopa (New Jersey):

    --2013-02-26 10:16:22--  http://198.211.100.5/test.dat
Connecting to 198.211.100.5:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 10485760 (10M) [application/octet-stream]
Saving to: â/dev/nullâ

100%[===========================================================================================================================>] 10,485,760  2.00M/s   in 5.2s

2013-02-26 10:16:27 (1.94 MB/s) - â/dev/nullâ
  • rds100rds100 Member

    Wasn't it unlimited bandwidth? What are you complaining about.

  • support123support123 Member

    @rds100 said: Wasn't it unlimited bandwidth? What are you complaining about.

    lol,They have now limit on transfer .But network is more oversold than 99% of the providers here

  • krypskryps Member

    @MiguelQ Maybe they got a direct pipe to you. I just measured again and performance is still abysmal (<100KB/sec) from OVH FR, BuyVM LAX, ...

  • MiguelQMiguelQ Member

    @kryps said: Maybe they got a direct pipe to you.

    I wouldn't call it a "direct pipe", but it is closer than OVH FR and BuyVM LV for sure:

     2  ethernet13-7-br1.pnj1.choopa.net (108.61.92.145)  11.233 ms  11.127 ms  10.886 ms
 3  ve67-br1.nyc2.choopa.net (108.61.93.110)  1.237 ms 108.61.1.38.choopa.net (108.61.1.38)  1.256 ms ve67-br1.nyc2.choopa.net (108.61.93.110)  1.130 ms
 4  bundle-ether2.nyktr1.NewYork.opentransit.net (193.251.255.185)  4.246 ms  4.017 ms bundle-ether1.nyktr1.NewYork.opentransit.net (193.251.251.101)  7.754 ms
 5  xe-5-0-0.edge2.NewYork1.Level3.net (4.68.111.81)  1.307 ms  1.240 ms  1.147 ms
 6  vlan90.csw4.NewYork1.Level3.net (4.69.155.254)  1.624 ms vlan80.csw3.NewYork1.Level3.net (4.69.155.190)  1.873 ms  1.610 ms
 7  ae-81-81.ebr1.NewYork1.Level3.net (4.69.134.73)  2.030 ms ae-61-61.ebr1.NewYork1.Level3.net (4.69.134.65)  1.745 ms ae-91-91.ebr1.NewYork1.Level3.net (4.69.134.77)  1.941 ms
 8  ae-2-2.ebr1.Newark1.Level3.net (4.69.132.98)  1.724 ms  1.604 ms  1.718 ms
 9  ae-1-51.edge3.Newark1.Level3.net (4.69.156.11)  1.634 ms  1.824 ms  1.762 ms
10  SERVERSTACK.edge3.Newark1.Level3.net (4.28.6.70)  2.338 ms  2.547 ms  2.418 ms
11  198.211.100.5 (198.211.100.5)  3.325 ms  3.156 ms  3.074 ms
  • Nick_ANick_A Member, Top Host, Host Rep

    @murky said: ITT: bitchy idiots who are too scared to tell support they're having bandwidth performance issues

    Yes, guys, God forbid you should talk about a provider's performance on a VPS forum. Come on.

  • murkymurky Member

    @Nick_A said: Yes, guys, God forbid you should talk about a provider's performance on a VPS forum. Come on.

    k

    RamNode doesn't give me 2000MB/s IO. :((((((

  • Nick_ANick_A Member, Top Host, Host Rep

    @murky said: RamNode doesn't give me 2000MB/s IO. :((((((

    You must be too scared to tell support I guess! Go open a ticket and don't ever post on a forum about performance!

  • CoreyCorey Member

    @Nick_A said: You must be too scared to tell support I guess! Go open a ticket and don't ever post on a forum about performance!

    @murky said: k

    RamNode doesn't give me 2000MB/s IO. :((((((

    Get a room ;)

  • IshaqIshaq Member

    @murky said: ITT: bitchy idiots who are too scared to tell support they're having bandwidth performance issues

    Please, get a grip.

    Haha. Now I'm almost certain you are gubbyte.

  • earlearl Member

    @Ishaq said: Haha. Now I'm almost certain you are gubbyte.

    I thought maybe he was subigo?

  • Nick_ANick_A Member, Top Host, Host Rep

    @Ishaq said: Haha. Now I'm almost certain you are gubbyte.

    Yes, must be.

  • support123support123 Member
    edited February 2013

    @Nick_A said: Yes, must be.

    Ramnode is more stable and predictable that DO :)

  • AmitzAmitz Member

    At least we know now. The DO fanboy has a new name...
    image

  • murkymurky Member

    @Amitz said: DO fanboy

    Funny guy, you are.

  • AmitzAmitz Member
    edited February 2013

    @murky said: Funny guy, you are.

    Sometimes.
    But we have something in common. I am triggered by you like you are triggered by DO. Time for me to meet the people at Betty Ford again. I was way funnier while still on drugs ;)

  • lumaluma Member

    From DO to Ramnode I get:
    2013-02-26 18:04:10 (19.3 MB/s) - `test100.zip' saved [104874307/104874307]

    DO to Ionity:
    2013-02-26 10:05:16 (19.0 MB/s) - `test100.zip' saved [104874307/104874307]

    So not sure why yours is so slow. Perhaps open a ticket?

  • krypskryps Member

    @luma Did you try their official DL links? Those are faster for me too. Please try with the download links I posted above.

  • lumaluma Member

    @kryps said: Did you try their official DL links? Those are faster for me too. Please try with the download links I posted above.

    I used my own VPS in New york with them. I grabbed the 100 meg test file from Softlayer and put it on my VPS at DO and then used a few other VPS' to test it.

  • unusedunused Member

    From 1gb on a 10gb uplink in NY:

    wget -O /dev/null http://198.211.100.5/test.dat 
--2013-02-26 13:26:30--  http://198.211.100.5/test.dat
Connecting to 198.211.100.5:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 10485760 (10M) [application/octet-stream]
Saving to: `/dev/null'

100%[====================>] 10,485,760  1.64M/s   in 5.8s    

2013-02-26 13:26:37 (1.71 MB/s) - `/dev/null' saved [10485760/10485760]

    What's surprising is this is pure cogent -> cogent.

    64 bytes from 198.211.100.5: icmp_seq=1 ttl=57 time=1.24 ms
64 bytes from 198.211.100.5: icmp_seq=2 ttl=57 time=1.55 ms

    Definitely DO is maxing out their pipe, or they are limiting your outbound bandwidth.

    And to the amsterdam:

     wget -O /dev/null http://198.211.124.199/test.dat 
--2013-02-26 13:26:44--  http://198.211.124.199/test.dat
Connecting to 198.211.124.199:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 10485760 (10M) [application/octet-stream]
Saving to: `/dev/null'

100%[====================>] 10,485,760  2.84M/s   in 3.5s    

2013-02-26 13:26:48 (2.84 MB/s) - `/dev/null' saved [10485760/10485760]

    Still, for the price can you complain?

  • lumaluma Member

    @unused said: Definitely DO is maxing out their pipe, or they are limiting your outbound bandwidth.

    Maybe his Node is maxed but not the network as I am consistently getting 15-20MB/s (see above post by me)

  • unusedunused Member

    @luma url of your test file?

  • lumaluma Member

    @unused said: @luma url of your test file?

    I will pm you (and anyone who wants it) but I don't want to paste the ip here in the forum)

  • unusedunused Member

    Sure - will be interesting to see the results vs @kryps ip - I'll just share the download rate

  • unusedunused Member

    Well, @luma is right - downloading his test file from DO in NY I get very different results:

    HTTP request sent, awaiting response... 200 OK
Length: 104874307 (100M) [application/zip]
Saving to: `test100.zip'

100%[=====================>] 104,874,307 30.0M/s   in 3.4s    

2013-02-26 18:36:19 (29.0 MB/s) - `test100.zip' saved [104874307/104874307]

    Ping is even a bit less

    icmp_seq=1 ttl=57 time=0.985 ms
icmp_seq=2 ttl=57 time=0.975 ms

    What's interesting is that the trace is identical to DO

     5  te4-2.ccr01.jfk08.atlas.cogentco.com (154.54.27.214)  1.343 ms  1.570 ms  1.441 ms
 6  38.122.106.2 (38.122.106.2)  0.981 ms  0.887 ms  0.974 ms
 7  69.55.60.17 (69.55.60.17)  1.024 ms  1.067 ms  0.949 ms
..... then the node

    So that does indicate some sort of node or at least switch specific issue with @kryps vm --- or perhaps they are actually limiting you for some reason.

Sign In or Register to comment.