We have 5 physical machines hosted at OVH. Each of them have Proxmox 5.2-10 installed
Machines are inside of OVH vRack.
we have separate LXC for database, separate LXC with loadbalancer and several separate instances with our web application stack (php and nginx).
When we put more LXCs with our web stack on a single physical machine after a while all of them except one just stuck.
PHP nor nginx are not responding on the network. Load average on the physical server suddenly starts to grow.
It is possible to enter the LXC via ssh or pct enter but impossible to shutdown from Proxmox web interface, LXC process has to be forcefully killed.
This problem started to appear when we upgraded PHP from 7.0 to 7.2. I don't think that it could be somehow related but I've never experienced simillar behaviour before.
So far we weren't able to come up with any solution except keeping the web LXCs on separate machines.
LXCs which do not have our web app stack can be normally together on one physical machine.
We believe that it has to do something with networking. Some files are shared via NFS on all LXCs. Could this be a hint? But again the problem is only on web LXCs. Other container with shared files can normaly work together.
Example configration of LXC network interfaces:
Example configuration of host machine interfaces:
Machines are inside of OVH vRack.
we have separate LXC for database, separate LXC with loadbalancer and several separate instances with our web application stack (php and nginx).
When we put more LXCs with our web stack on a single physical machine after a while all of them except one just stuck.
PHP nor nginx are not responding on the network. Load average on the physical server suddenly starts to grow.
It is possible to enter the LXC via ssh or pct enter but impossible to shutdown from Proxmox web interface, LXC process has to be forcefully killed.
This problem started to appear when we upgraded PHP from 7.0 to 7.2. I don't think that it could be somehow related but I've never experienced simillar behaviour before.
So far we weren't able to come up with any solution except keeping the web LXCs on separate machines.
LXCs which do not have our web app stack can be normally together on one physical machine.
We believe that it has to do something with networking. Some files are shared via NFS on all LXCs. Could this be a hint? But again the problem is only on web LXCs. Other container with shared files can normaly work together.
Example configration of LXC network interfaces:
Code:
auto lo
iface lo inet loopback
iface lo inet6 loopback
# interfaces(5) file used by ifup(8) and ifdown(8)
# Include files from /etc/network/interfaces.d:
source-directory /etc/network/interfaces.d
auto eth0
iface eth0 inet static
address 5.39.56.172
netmask 255.255.255.224
gateway 5.39.56.190
auto eth1
iface eth1 inet static
address 172.16.32.172
netmask 255.240.0.0Example configuration of host machine interfaces:
Code:
auto lo
iface lo inet loopback
iface eno1 inet manual
auto eno2
iface eno2 inet manual
auto vmbr0
iface vmbr0 inet static
address 147.135.254.187
netmask 255.255.255.0
gateway 147.135.254.254
bridge-ports eno1
bridge-stp off
bridge-fd 0
iface vmbr0 inet6 static
address 2001:41d0:303:2dbb::
netmask 64
auto vmbr1
iface vmbr1 inet static
address 172.16.35.1
netmask 255.240.0.0
broadcast 172.31.255.255
bridge-ports eno2
bridge-stp off
bridge-fd 0
network 172.16.0.0