Proble with Centos 5 When upgrading from 6.4 to 7. X and 8.1.

[Silverstein]

I had a Centos 5 VM running fine on a VM. under PVE 6.4. I upgraded to PVE 7, and it started being very slow and having network connectivity between the VM and the local physical network dropping and restoring on its own. When I used the original of the VM on another system which had a PVE 8 system installed new, it is seems happy.

I was thinking it may have been an issue with the upgrade not being quite right. Or is there an issue with Centos 5 and modern versions of PVE?

I tried both the intel and the virtio for the network device.

Thank you.

 

[LnxBil]

Or is there an issue with Centos 5 and modern versions of PVE?

That OS went EOL 7 years ago. There could be a lot of things wrong.

My experience is that older OS run much slower in general than any modern OS due to the lack of virtio support. The Hypervisor and the Guest needs to implement a driver and that will not speedup things. Sometimes, Virtualizing another Hypervisor in order to virtualize the older guest is faster than running it directly in PVE. I haven't run anything RHEL5 based as a VM for many years, so I really cannot say.

In order to have proper speed, what about running it inside of an LX(C) container if RHEL5-based images are still supported there. Maybe even migrate to a Docker container.

 

[Silverstein]

I know it went Centos went EOL. It also keeps dropping the network connection to the local network. It works fine on a fresh 8.1 install. It was working fine on the old PVE 6.4. But the fresh 8.1 is not going directly to the network, but through a pfsense. Would I be better off trying a fresh 8.1 install, or reinstall 6.4?

 

[LnxBil]

I don't think that the problem lies on the hypervisor. Have you tried looking at the logs in the centos container? Does the container use DHCP or is the IP hardcoded? Last resort is always running tcpdump and analyzing what went wrong.

 

[Silverstein]

I don't think that the problem lies on the hypervisor. Have you tried looking at the logs in the centos container? Does the container use DHCP or is the IP hardcoded? Last resort is always running tcpdump and analyzing what went wrong.

I wiped the system, did a fresh install of PVE 8.1, and restored the VM. It's happy now.