You need to log in to create posts and topics. Login · Register

Failed Dependency on setting up network

Pages: 1 2 3

For status of ifup@eth2.7

Main PID: 3716 (code=exited, status=1/FAILURE)
ifup: failed to bring up eth2.7
ifup@eth2.7.service: Main process exited, code=exited, status=1/FAILURE
ifup@eth2.7.service: Failed with result 'exit-code'.

For status of systemd-networkd: seems to be disabled.

Active: inactive (dead)

For status of networking:

Active: active (exited)
Main PID: 3327 (code=exited, status=0/SUCCESS)

Do you think I can disable this service to avoid startup problems with the following command?

systemctl disable ifup@eth2.7

Or can it be removed so I don't have to work about it getting activated again?

Yes you can disable it. it is not used or enabled by PetaSAN, i checked on some installed systems and it is not enabled.

I tried to disable, but I got back a blank and the service is still loaded.
How is that service registered during installation?
How can I disable this?

Had to use this command instead.

systemctl reset-failed ifup@eth2.7.service

On start up though, this service came back.
This must be called from another startup script somewhere.
This is still causing start up issues.
Is this some sort of legacy way of starting up interfaces?

Seems to be a deeper problem. This service seems to be loaded by networking in the ifupdown.service as a preload.
It might be startup order. Meaning that if udev has not renamed the interface by the time this service starts, the ifupdown service cannot find eth2.7 and will fail.
This is a working theory, and I am researching more.

Any help will be appreciated.

if you can reproduce in another environment, please post the steps and we will try it in our cluster.

I will try to replicate this week.

Hi podilarius!

Any news on this matter? I have the same problem with version 3.2.1, the management is on VLAN 1254 and I keep installing node-1 adding to cluster and as it boots doesn't have connectivy until I do a systemctl networking restart, the networking starts to work but doesn't get available as a node.

Is there a way to restart the the networking and the cluster services? At least for now as I don't figure out whats going on?

Update:

After running /opt/petasan/scripts/start_petasan_services.py and a reboot... magically it started working without the error... later today I will reboot node-1 to see if it keeps working or it fails again...

Pages: 1 2 3