Networking Host/Guest – Bridged Networking – Virtual Networking – Issues

[Server.Experts]

Assistance required with >>> networking <<< setup :
CentOS 5 OpenVZ – Container
kernel - 2.6.18-2-pve from ProxMox-VE v1.5

# experiencing difficulty setting up bridged network
# the container fails to connect to external network after boot

 

[Server.Experts]

Assistance required with >>> networking <<< setup :
CentOS 5 OpenVZ – Container
kernel - 2.6.18-2-pve from ProxMox-VE v1.5

# experiencing difficulty setting up bridged network
# the container fails to connect to external network after boot

All assistance and guidance would be much appreciated , Thank You

 

[Server.Experts]

Additionally it has been found that the host network fails intermittently, as well as well as NODE to MASTER and MASTER to NODE authentication / synchronization fails
# screenshot 1 shows the nodes, error codes and cluster uptime, issue #1
# screenshot 2 shows the nodes, error codes and cluster uptime, issue #2

Additionally, several different test setups were tried and similar problem was recreated

All nodes/servers running latest Proxmox VE distribution

pls see attachments

 

[Server.Experts]

Additionally it has been found that the host network fails intermittently, as well as well as NODE to MASTER and MASTER to NODE authentication / synchronization fails
# screenshot 1 shows the nodes, error codes and cluster uptime, issue #1
# screenshot 2 shows the nodes, error codes and cluster uptime, issue #2

Additionally, several different test setups were tried and similar problem was recreated

All nodes/servers running latest Proxmox VE distribution

pls see attachments

It has also been noted that within this state (i.e. non-synchronized & non-authorized)
# (OpenVZ containers) can be MIGRATED between MASTER to NODE, but remain un-operatable, and can be transferred back
# Although containers can be migrated between Node to Node, it becomes a one way operation, they become unrecoverable and inaccessible after migration

#video exist to show this issue, but due to size restricted to upload it…

 

[dietmar]

Do you have correct time on all nodes?

 

[Server.Experts]

I would verify and get back to you on this Thanks

 

[Server.Experts]

Do you have correct time on all nodes?

system clocks are in sync and are accurate to minutes (but not seconds).

additionally on multiple boots the nodes come to sync ...... and intermittently go offline .... this results in VM's loosing network as well .... untill correct reboot is achieved .... I do apologize dietmar, but I have not been able to find a pattern to this problem.

I await to hear from you. Thanks

 

[Server.Experts]

Do you have correct time on all nodes?

dietmar

• View Profile
• View Forum Posts
• Add as Contact

please do advise what the potential problem could be. Thanks

 

[dietmar]

additionally on multiple boots the nodes come to sync ...... and intermittently go offline .... this results in VM's loosing network as well .... untill correct reboot is achieved .... I do apologize dietmar, but I have not been able to find a pattern to this problem.

Does the network work on the host?

 

[Server.Experts]

Does the network work on the host?

Hi Dietmar,
Unfortunately, yes / no, the answer is in line with the persistent but yet intermittent nature of the issue.

 

[dietmar]

Hi Dietmar,
Unfortunately, yes / no, the answer is in line with the persistent but yet intermittent nature of the issue.

Please check if you have duplicate IP or MAC addresses somewhere.

If notheing helps, maybe you can test with another NIC. Also check your network equipement (switch).

 

[BiagioParuolo]

I'm too, I had this problem.
My problem was duplicate MAC address on openvz net card. Then I delete net card into VM config and reinstall.

 

[Server.Experts]

Please check if you have duplicate IP or MAC addresses somewhere.

If notheing helps, maybe you can test with another NIC. Also check your network equipement (switch).

Dietmar you are SuperStar !!! Please remind me to buy you a bear !!!


for everyone else who experiences similar issue...
• for mass duplication and deployment of OpenVZ containers
• please ensure your script contains necessary parameters for incrementally omitting/changing MAC address(es) inside containers ( applicable to bridged [VETH]connections)
• it can also be done manually but would be tedious for larger deployment
• if the above instructions are not practiced/implemented, you are likely to experience networking issues later on, which may include random connection issues, disconnections, packet transmission failures etc
• additionally system clocks need to be accurately and precisely set in all the clusters, even down to correct time-zones, I believe, this is to prevent sync failures/timeouts


Thank You to all the developers, contributorsand sponsorsof this project, we sincerely appreciate your efforts, hard-work and continued support


Additionally, Dietmar could you update us on the integration status of LXC into the new version of ProxMox and how we can assist. As you might be aware the OpenVZ developers have indicated it is likely to supersede their project, due to it being part of the mainline kernel.