Howdy, Stranger!

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


Kimsufi IPV6 connection drops on VMs
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.

Kimsufi IPV6 connection drops on VMs

v3ngv3ng Member, Patron Provider
edited September 2021 in Help

Hey there,

I've ordered a Kimsufi KS-7 in BHS a few days ago which I'd like to use it as a backup server and run Proxmox on it.

However my Debian 11 VMs keep loosing the IPv6 connection after some time and I've got no idea why.

The host itself seems to keep the IPv6 connection however.

I've attached my network, ndppd, sysctl and radvd config below.

Thank you!

#/etc/ndppd.conf

route-ttl 30000
  proxy eth0 {
   router no
   timeout 500
   ttl 30000
   rule 2607:xxxx:60:a57::/64 {
    static
   }
  }
# /etc/sysctl.conf

net.ipv6.conf.all.proxy_ndp = 1
net.ipv6.bindv6only = 0
net.ipv6.conf.all.forwarding=1
net.ipv6.conf.all.autoconf=0
net.ipv6.conf.all.accept_ra=0
# /etc/radvd.conf
interface vmbr0 {
  AdvSendAdvert on;
  prefix 2607:xxxx:60:a57::/64
  {
    AdvOnLink on;
    AdvAutonomous on;
  };
RDNSS 2606:4700:4700::1111 {
        AdvRDNSSLifetime 3600;
 };
};
# network interfaces
 auto lo
   iface lo inet loopback


 auto eth0
   iface eth0 inet dhcp

    iface eth0 inet6 static
    address 2607:xxxx:60:a57::2
    netmask 128
    post-up /sbin/ip -f inet6 route add 2607:xxxx:60:aFF:FF:FF:FF:FF dev eth0 
    post-up /sbin/ip -f inet6 route add default via 2607: xxxx:60:aFF:FF:FF:FF:FF 
    pre-down /sbin/ip -f inet6 route del 2607: xxxx:60:aFF:FF:FF:FF:FF dev eth0
    pre-down /sbin/ip -f inet6 route del default via 2607: xxxx:60:aFF:FF:FF:FF:FF


  auto vmbr0
    iface vmbr0 inet6 static
    address 2607: xxxx:60:a57::2
    netmask 64
    bridge_ports none
    bridge_stp off
    bridge_fd 0
    bridge_maxwait 0
    post-up echo 1 > /proc/sys/net/ipv6/conf/all/proxy_ndp
    post-up echo 1 > /proc/sys/net/ipv6/conf/all/forwarding
    post-up echo 1 > /proc/sys/net/ipv6/conf/default/forwarding

Comments

  • NeoonNeoon Community Contributor, Veteran
    edited September 2021

    Try ndp proxy and configure it static.

    https://wiki.x8e.net/doku.php?id=proxmox_ipv6

  • v3ngv3ng Member, Patron Provider

    Thanks, I just configured static IPs on the VMs and will see how it goes.
    But I hope it‘s not related to radvd since I use it for my IP deployment

  • v3ngv3ng Member, Patron Provider

    Nope still the same issue

  • v3ngv3ng Member, Patron Provider

    It seems to help if you ping the VM and wait for some time, then the IPv6 connection seems to come back

  • NeoonNeoon Community Contributor, Veteran
    edited September 2021

    ip -6 neighbour

    Check if ndp entries are REACHABLE and not STALE, especially the entry to the router.

  • v3ngv3ng Member, Patron Provider
    edited September 2021

    Quite a few are listed as STALE.

    However I just tried connecting to one of the STALE IPs which worked just fine and the status changes to REACHABLE after the connection.

    Btw: I've disabled the IPv6 privacy extension on the VM a few minutes ago, we'll see if that helps.

    cat >> /etc/sysctl.conf <<EOT
    net.ipv6.conf.all.use_tempaddr=0
    net.ipv6.conf.default.use_tempaddr=0
    EOT
    
    sysctl -p
    

    Edit: Disabling the privacy extension didn’t help unfortunately

  • v3ngv3ng Member, Patron Provider

    @Neoon said:
    ip -6 neighbour

    Check if ndp entries are REACHABLE and not STALE, especially the entry to the router.

    On the VM itself it only shows stale, even the entry to the router

  • rm_rm_ IPv6 Advocate, Veteran

    What about firewall on the host or VM? Any rules?

  • v3ngv3ng Member, Patron Provider

    No, nothing

  • v3ngv3ng Member, Patron Provider

    Anyone got an idea?

  • rm_rm_ IPv6 Advocate, Veteran

    @v3ng said: Anyone got an idea?

    Not really. Get tcpdump and start debugging: what does VPS send when it works and when not, what does the host receive, what does host send towards OVH in both cases.

    Also you didn't specify, can you ping VM from the host, when it loses connection to the Internet? And by chance does pinging restore it? Also try the other way, ping host from the VM.

  • v3ngv3ng Member, Patron Provider

    ping does seem to restore it, I'll do some more testing once the connection drops again

  • v3ngv3ng Member, Patron Provider
    edited September 2021

    Alright I’m now 99% sure it’s some issue on OVH‘s side.

    I reinstalled the whole server again, checked every config and monitored the VMs.

    Was quite stable for a few hours and then it began to drop every now and then.

    Even the Host itself lost its IPv4 link a few hours ago….

    Looks like the interface (vmbr0) somehow dropped, but I‘ve got absolutely no idea why all of this is happening with the OVH server.
    Maybe a broken Proxmox image? Even tho I updated it to the latest Proxmox 7 version.


    ntpd[1064]: Deleting interface #3 eth0, 198.xxx.147.87#123, interface stats: received=972, sent=942, dropped=0, active_time=86396 secs

    https://bugs.launchpad.net/ubuntu/+source/ntp/+bug/1069543

  • NeoonNeoon Community Contributor, Veteran

    https://lala.im/8145.html

    Just leaving it here, put that same setup on my KS-LE, will report how stable it is.

  • @Neoon said:
    https://lala.im/8145.html

    Just leaving it here, put that same setup on my KS-LE, will report how stable it is.

    Need @yoursunny to translate :)

    @v3ng said:
    ping does seem to restore it, I'll do some more testing once the connection drops again

    Correct. From my *BSD host, I used ndp -a and saw that TTL is 24h, if after 24h without imcp6 connection from host to guest, connection will be lost. I tried many methods or tweaks, but no difference, because issues come from the way SYS/KS configured.

    @v3ng said:
    Alright I’m now 99% sure it’s some issue on OVH‘s side.

    No, incorrect. It's 100%.
    Dont waste your time, just write a cron script to ping -6 from host to every guest inside it (within 24h period).

  • NeoonNeoon Community Contributor, Veteran

    @bsh said:

    @Neoon said:
    https://lala.im/8145.html

    Just leaving it here, put that same setup on my KS-LE, will report how stable it is.

    Need @yoursunny to translate :)

    You don't need even to speak Large Plate of Chicken to understand that the manuale.

  • :) Yep ! I could read image text. I think it would work, because he/she proposed to set Min/Max for retrieval in a short time 3 and 10, and the range is within 24h, even that guest VM is idle.

  • v3ngv3ng Member, Patron Provider

    @bsh said:

    @Neoon said:
    https://lala.im/8145.html

    Just leaving it here, put that same setup on my KS-LE, will report how stable it is.

    Need @yoursunny to translate :)

    @v3ng said:
    ping does seem to restore it, I'll do some more testing once the connection drops again

    Correct. From my *BSD host, I used ndp -a and saw that TTL is 24h, if after 24h without imcp6 connection from host to guest, connection will be lost. I tried many methods or tweaks, but no difference, because issues come from the way SYS/KS configured.

    @v3ng said:
    Alright I’m now 99% sure it’s some issue on OVH‘s side.

    No, incorrect. It's 100%.
    Dont waste your time, just write a cron script to ping -6 from host to every guest inside it (within 24h period).

    It somehow works on my SYS box in SBG, even tho the setup is the same

  • dev_vpsdev_vps Member
    edited December 2021

    Can someone point me to steps to install proxmox 7?

    I have heard upgrading proxmox from 6.x to 7 is problematic and can lose network access.

    Thanks so much

  • NeoonNeoon Community Contributor, Veteran

    @dev_vps said:
    Can someone point me to steps to install proxmox 7?

    I have heard upgrading proxmox from 6.x to 7 is problematic and can lose network access.

    Thanks so much

    Proxmox 6 Template => fucked
    Proxmox 6 to 7 => fine, just need to add hwaddress
    Proxmox 7 install on Debian 11 => fucks the network config entirely, you need to make a backup and restore before reboot plus apply hwaddress

Sign In or Register to comment.