ha-cluster script return error for default network interface

SLES High Availability Extension
1

Hello SME,

I’m testing SLES12 SP3 HA Extension. When running ha-cluster-init in node 1, it returns error message as follows. From the man page, it mentions eth0 is the default network interface. Don’t understand why it cannot detect the network card. The workaround is to use -i option to specify eth0 and looks worked fine. When running ha-cluster-join in node 2, it returns the same error message. But this time, that workaround doesn’t work and it returns another error message as follows. I verified that node 1 and node 2 are able to login with each others via ssh without password. Do you have any idea to solve it?

In node 1

ha-cluster-init
“ERROR: cluster.init: Failed to determine default network interface”

hc-cluster-init -i eth0
…
…
Done (log saved to /var/log/ha-cluster-bootstrap.log)

In node 2

ha-cluster-join
ERROR: cluster.join: Failed to determine default network interface

ha-cluster-join -i eth0
…
…
ERROR: 1: cluster.join: Can’t invoke ha-cluster-init ssh_remote on 192.168.10.11

2

Hi collinsyunkk,

had you checked what’s in the logs (i. e. /var/log/ha-cluster-bootstrap.log) when running cluster-init without “-i”? Also, does “ip a” report anything unusual?

Regards,
J

3

[QUOTE=jmozdzen;55116]Hi collinsyunkk,

had you checked what’s in the logs (i. e. /var/log/ha-cluster-bootstrap.log) when running cluster-init without “-i”? Also, does “ip a” report anything unusual?

Regards,
J[/QUOTE]

Hello jmozdzen,

Thanks for your comment. After read the log, I got a hint to turn on debug mode when running the crm command which called by ha-cluster-join. The issue is fixed by adding default gateway in both node 1 and 2 (even if both nodes are located at the same subnet).

Regards,
Collins

4

Hi Collins,

thank you for sharing your findings - I wouldn’t have put my guess on a missing default route, indeed. Glad you found that out and got it running!

Regards,
J