OpenVZ VPS Crashes during Speed test

Why would an OpenVZ VPS crash by doing a speed test when using Shadowsocks?

Tagged:

Comments

  • What does crash mean?
    Please post error message (in textual format, not pictures).

  • Crash means the end.

    Therefore, OP means the end is nigh.

    Thanked by (1)lindy54

    ♻ Amitz day is October 21.
    ♻ Join Nigh sect by adopting my avatar. Let us spread the joys of the end.

  • edited March 2022

    @yoursunny said:
    What does crash mean?
    Please post error message (in textual format, not pictures).

    VM shuts down, and rebooting it is not possible from the provider's control panel for a while (from couple of hours to days).

    Thanked by (1)ceddr
  • And the problem is not just about shadowsocks, It happens with wireguard as well.> @lindy54 said:

    Why would an OpenVZ VPS crash by doing a speed test when using Shadowsocks?

    Thanked by (2)amiir ceddr
  • How much memory does your OpenVZ container have?

    Thanked by (1)Not_Oles

    Cheap dedis are my drug, and I'm too far gone to turn back.

  • @CamoYoshi said:
    How much memory does your OpenVZ container have?

    256 MB

    Thanked by (2)amiir ceddr
  • Well, there is your answer right there.

    ♻ Amitz day is October 21.
    ♻ Join Nigh sect by adopting my avatar. Let us spread the joys of the end.

  • edited March 2022

    @deank said:
    Well, there is your answer right there.

    But Openvpn worked without such issues, so low ram is a problem with shadowsocks and wireguard for just one client?

    Thanked by (1)amiir
  • The simplest solution for the answer would be do the same task but with half a gig of RAM.
    If it doesn't scream the end, you know the culprit.

    ♻ Amitz day is October 21.
    ♻ Join Nigh sect by adopting my avatar. Let us spread the joys of the end.

  • edited March 2022

    @lindy54 said:

    @deank said:
    Well, there is your answer right there.

    But Openvpn worked without such issues, so low ram is a problem with shadowsocks and wireguard for just one client?

    Shadowsocks and Wireguard-go need more RAM compared to the OpenVPN daemon. Simple as that. The more recent versions of each upped their RAM requirements. 256MB is not sufficient to run a system + the two clients, unless you really slim down the underlying system processes, use some swap space, and only run either Shadowsocks or Wireguard-go, but not both at the same time.

    Likely the reason it's crashing is due to OOM killer killing a needed system process which causes the container to halt. When you don't have a SWAP to fail back onto, the kernel has to resort to OOM killer to not run out of RAM, lest the entire thing comes crashing down (which is likely what's happening here).

    Thanked by (1)Not_Oles

    Cheap dedis are my drug, and I'm too far gone to turn back.

  • FrankZFrankZ Moderator

    @lindy54 said: VM shuts down, and rebooting it is not possible from the provider's control panel for a while (from couple of hours to days).

    The fact that the VPS is unable to start again for

    (from couple of hours to days)

    suggests that there may be an abuse script on the providers side that is causing this.
    Care to name the provider?

    Thanked by (3)Not_Oles skorous lindy54

    For staff assistance or support issues please use the helpdesk ticket system at https://support.lowendspirit.com/index.php?a=add

  • Yeah, name, shame, and then logically file a lawsuit for emotional damages.

    ♻ Amitz day is October 21.
    ♻ Join Nigh sect by adopting my avatar. Let us spread the joys of the end.

  • vyasvyas OGSenpai
    edited March 2022

    OpenVZ

    Which version? Which OS?

  • edited March 2022

    v2ray and wireguard-go work fine in webhorizon NAT-256 vps.

    https://www.speedtest.net/result/c/2850b824-da4b-421e-b9b8-46c60539c3f1

    Thanked by (1)CamoYoshi
  • @FrankZ said:

    @lindy54 said: VM shuts down, and rebooting it is not possible from the provider's control panel for a while (from couple of hours to days).

    The fact that the VPS is unable to start again for

    (from couple of hours to days)

    suggests that there may be an abuse script on the providers side that is causing this.
    Care to name the provider?

    Yes, I also suspect that the problem may be caused by an abuse script. The funny thing is that the main ipv4 address gets inaccessible for that period too (from different IPs, so it's not an IP blocking mechanism). I think there will be some issues if I name the provider.

    Thanked by (1)FrankZ
  • @vyas said:

    OpenVZ

    Which version? Which OS?

    OpenVZ 7, happens on both CentOS 7 and Ubuntu 18.04.

  • @feie said:
    v2ray and wireguard-go work fine in webhorizon NAT-256 vps.

    https://www.speedtest.net/result/c/2850b824-da4b-421e-b9b8-46c60539c3f1

    But I have the same issue with v2ray unfortunately.

    Thanked by (1)amiir
  • So, I removed wireguard completely and shadowsocks works well for speeds above 100 mbps without "crashing".
    I doubt that it's a ram issue though, because shadowsocks continues to run ok under much higher ram loads.

    Thanked by (1)amiir
  • @lindy54 said:
    So, I removed wireguard completely and shadowsocks works well for speeds above 100 mbps without "crashing".
    I doubt that it's a ram issue though, because shadowsocks continues to run ok under much higher ram loads.

    wireguard-go, of course.

    Thanked by (1)amiir
  • mikhomikho AdministratorOG

    @lindy54 said: I think there will be some issues if I name the provider.

    Have you opened a ticket with the provider regarding this?

    “Technology is best when it brings people together.” – Matt Mullenweg

  • @mikho said:

    @lindy54 said: I think there will be some issues if I name the provider.

    Have you opened a ticket with the provider regarding this?

    No, because I couldn't determine what exactly causes it.

    Thanked by (2)amiir ceddr
  • mikhomikho AdministratorOG

    @lindy54 said:

    @mikho said:

    @lindy54 said: I think there will be some issues if I name the provider.

    Have you opened a ticket with the provider regarding this?

    No, because I couldn't determine what exactly causes it.

    Open a ticket with the provider. They might help find the reason as they have access to things you dont.

    Thanked by (1)lindy54

    “Technology is best when it brings people together.” – Matt Mullenweg

This discussion has been closed.