Howdy, Stranger!

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


Slow VPS Download speeds till I reboot
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.

Slow VPS Download speeds till I reboot

HuntersPadHuntersPad Member
edited October 2016 in Help

This has really stumped me.
Running CentOS 6.5 with plesk

Disk I/O is fine
Network is fine.

Its on an OVH server in Canada from a Reseller.

Example: uploading via FTP or Downloading it via Wget speeds are fine around 20MiB/s so perfect. After about 10 or so mins speeds craw to around 200-200kbps! Until I reboot, when I reboot speeds are instantly perfect again! Is there something to do with CentOS that I am missing? I don't think its the provider as he is not having this issue on his side.

This VPS is not live website wise so its not a traffic issue.

Any suggestions?

Would I be better off switching to Debian with Plesk to try out?

Thanks.
Here is top while speeds have slowed down again after 11 mins.

top - 13:06:39 up 11 min, 1 user, load average: 0.02, 0.14, 0.14

Tasks: 208 total, 1 running, 207 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 1921348k total, 501972k used, 1419376k free, 18104k buffers
Swap: 2097148k total, 0k used, 2097148k free, 282564k cached

Meant to add as well, the VPS upload speeds are fine. Its just the download speeds

Comments

  • What's top look like while you're trying to download? Linux buffers downloads in memory until exhausted, and then starts writing to disk concurrently. If the disk system that the VM is running on is overloaded, your iowait will spike while doing so.

  • HuntersPadHuntersPad Member
    edited October 2016

    @Damian said:
    What's top look like while you're trying to download? Linux buffers downloads in memory until exhausted, and then starts writing to disk concurrently. If the disk system that the VM is running on is overloaded, your iowait will spike while doing so.

    top - 13:42:59 up 26 min, 1 user, load average: 0.02, 0.08, 0.08
    Tasks: 210 total, 1 running, 209 sleeping, 0 stopped, 0 zombie
    Cpu(s): 0.0%us, 0.0%sy, 0.0%ni, 99.7%id, 0.3%wa, 0.0%hi, 0.0%si, 0.0%st
    Mem: 1921348k total, 436040k used, 1485308k free, 19168k buffers
    Swap: 2097148k total, 0k used, 2097148k free, 220720k cache

  • This sounds to me a lot more like a network issue where you are hitting a throttling point on your ISPs more expensive bandwidth segment. Where if you use the bandwidth to a more expensive route they QOS you after a short period. Have you tried running an MTR during your transfer to see if you experience any packet loss when this occurs? Also, can you provide a traceroute or a mtr report (blacking out your IP of course as needed) so we can see what the route actually looks like? It also sounds like rebooting your computer just so happens to take long enough for their QOS to reset, so when your back up its magically going fast again, until you hit that throttle point....

    Also, I would suggest if you haven't already making some adjustment to the sysctl.conf: this link provide some options for tweaking

    Let us know your findings once you tweak things a bit as well. Those options provided for KVM should generally translate for you.

    Also use paste.ee or a similar service to include your MTR or traceroute so we can see and advise on what may possibly be going on.

    Cheers!

    Thanked by 1HuntersPad
  • HuntersPadHuntersPad Member
    edited October 2016

    @TheLinuxBug said:
    This sounds to me a lot more like a network issue where you are hitting a throttling point on your ISPs more expensive bandwidth segment. Where if you use the bandwidth to a more expensive route they QOS you after a short period. Have you tried running an MTR during your transfer to see if you experience any packet loss when this occurs? Also, can you provide a traceroute or a mtr report (blacking out your IP of course as needed) so we can see what the route actually looks like? It also sounds like rebooting your computer just so happens to take long enough for their QOS to reset, so when your back up its magically going fast again, until you hit that throttle point....

    Also, I would suggest if you haven't already making some adjustment to the sysctl.conf: this link provide some options for tweaking

    Let us know your findings once you tweak things a bit as well. Those options provided for KVM should generally translate for you.

    Also use paste.ee or a similar service to include your MTR or traceroute so we can see and advise on what may possibly be going on.

    Cheers!
    @TheLinuxBug said:
    This sounds to me a lot more like a network issue where you are hitting a throttling point on your ISPs more expensive bandwidth segment. Where if you use the bandwidth to a more expensive route they QOS you after a short period. Have you tried running an MTR during your transfer to see if you experience any packet loss when this occurs? Also, can you provide a traceroute or a mtr report (blacking out your IP of course as needed) so we can see what the route actually looks like? It also sounds like rebooting your computer just so happens to take long enough for their QOS to reset, so when your back up its magically going fast again, until you hit that throttle point....

    Also, I would suggest if you haven't already making some adjustment to the sysctl.conf: this link provide some options for tweaking

    Let us know your findings once you tweak things a bit as well. Those options provided for KVM should generally translate for you.

    Also use paste.ee or a similar service to include your MTR or traceroute so we can see and advise on what may possibly be going on.

    Cheers!
    @TheLinuxBug said:
    This sounds to me a lot more like a network issue where you are hitting a throttling point on your ISPs more expensive bandwidth segment. Where if you use the bandwidth to a more expensive route they QOS you after a short period. Have you tried running an MTR during your transfer to see if you experience any packet loss when this occurs? Also, can you provide a traceroute or a mtr report (blacking out your IP of course as needed) so we can see what the route actually looks like? It also sounds like rebooting your computer just so happens to take long enough for their QOS to reset, so when your back up its magically going fast again, until you hit that throttle point....

    Also, I would suggest if you haven't already making some adjustment to the sysctl.conf: this link provide some options for tweaking

    Let us know your findings once you tweak things a bit as well. Those options provided for KVM should generally translate for you.

    Also use paste.ee or a similar service to include your MTR or traceroute so we can see and advise on what may possibly be going on.

    Cheers!

    Thanks for the info! I'll do more testing when I get back home. But I don't think it's throttling as a reboot brings it back to normal which only takes less than a minuet. Also I wouldnt see them needing to throttle download speeds on the vps, my upload speed stays perfectly normal It's also happens exactly at 10 min booted

  • TheLinuxBugTheLinuxBug Member
    edited October 2016

    hyelton said: Thanks for the info! I'll do more testing when I get back home. But I don't think it's throttling as a reboot brings it back to normal which only takes less than a minuet. Also I wouldnt see them needing to throttle download speeds on the vps, my upload speed stays perfectly normal It's also happens exactly at 10 min booted

    Quote

    Ahh, you mean when you reboot the VPS it works faster again? I though you meant you were downloading to your home connection and you were seeing that behavior. If it is indeed happening when you reboot the VPS, then what was said above with making those sysctl setting changes may help.

    let us know how your testing goes.

    Cheers!

  • @TheLinuxBug said:

    hyelton said: Thanks for the info! I'll do more testing when I get back home. But I don't think it's throttling as a reboot brings it back to normal which only takes less than a minuet. Also I wouldnt see them needing to throttle download speeds on the vps, my upload speed stays perfectly normal It's also happens exactly at 10 min booted

    Quote

    Ahh, you mean when you reboot the VPS it works faster again? I though you meant you were downloading to your home connection and you were seeing that behavior. If it is indeed happening when you reboot the VPS, then what was said above with making those sysctl setting changes may help.

    let us know how your testing goes.

    Cheers!

    Yeah Downloading on the VPS itself is throttled with everything until reboot.

    I hope I am not jinxing myself here, But I loaded Debian 7.3 and for 25 min the speed stayed above 15Mbps without issue. So now installing Plesk to see if thats causing the issue. If not then something with the CentOS repo was not going correctly.

  • @TheLinuxBug said:

    hyelton said: Thanks for the info! I'll do more testing when I get back home. But I don't think it's throttling as a reboot brings it back to normal which only takes less than a minuet. Also I wouldnt see them needing to throttle download speeds on the vps, my upload speed stays perfectly normal It's also happens exactly at 10 min booted

    Quote

    Ahh, you mean when you reboot the VPS it works faster again? I though you meant you were downloading to your home connection and you were seeing that behavior. If it is indeed happening when you reboot the VPS, then what was said above with making those sysctl setting changes may help.

    let us know how your testing goes.

    Cheers!

    Made those changes and still back to what is was. I guess I was just luckly earlier, but its now back to the same.

  • MikePTMikePT Moderator, Patron Provider, Veteran

    Tell your provider to disable ddos protection on your ip, then download and a few minutes later ask them to enable ddos protection on your ip.
    Note as for enabling ddos protection I mean always on (sorry for the typos, cellphone)

    At last but not least tell your provider to contact OVH. Its their throotling working. I've seen it before and took a couple of weeks to get them to fix it but they did.

  • MikeAMikeA Member, Patron Provider

    @MrGeneral said:
    Tell your provider to disable ddos protection on your ip, then download and a few minutes later ask them to enable ddos protection on your ip.
    Note as for enabling ddos protection I mean always on (sorry for the typos, cellphone)

    At last but not least tell your provider to contact OVH. Its their throotling working. I've seen it before and took a couple of weeks to get them to fix it but they did.

    Throttling? I've never had a problem like this with OVH, not on their main brand servers.

  • Update on the situation, It seems this slow speed after awhile on the VPS downstream side of things only slows down when going from Canada (where the VPS is) to Europe (At least the netherlands anyway) But reboot speeds things up for awhile and then slows back down. (Maybe it is ddos protection?) BUT speeds are normal and acceptable when downloading from US servers.

Sign In or Register to comment.