Change ipmitool timeouts in undercloud ironic (RHOSP 16.1)

Latest response

Hi,

the ironic conductor triggers an ipmitool command almost every 1 sec since the respond from the server's BMC takes "longer" than 1 sec. I took a look in the /var/log/containers/ironic/ironic-conductor.log and can see the ipmitool command issued:

....
Command: ipmitool -I lanplus -H XX.XX.XX.XX -L ADMINISTRATOR -p 623 -U admin -R 1 -N 1 -f /tmp/tmpa3l0iz8_ power status
Exit code: 1
....

The problem here is the -R & -N parameters, which are 1 sec here. If I execute the same ipmitool from the command line against the same server's BMC then I get also an error. If I increase the -N to "2" and execute the following ipmitool command from CLI:

ipmitool -I lanplus -H XX.XX.XX.XX -L ADMINISTRATOR -p 623 -U admin -R 1 -N 2 -P XXXXXXXX power status

then I get the power state returned and exit code equal 0.

I logged in the ironic_conductor container:

sudo podman exec -ti ironic_conductor /bin/bash

and added in [ipmi] within the /etc/ironic/ironic.conf:
retry_timeout = 2

I was expecting that the retry_timeout (as described in https://docs.openstack.org/ironic/5.1.3/_modules/ironic/drivers/modules/ipmitool.html) would reflect the -N parameter of the ipmitool triggered by the ironic-conductor. But that was not the case...

My question...how can I permanently change the the -R & -N parameters of the ipmitool command that is triggered by ironic in the undercloud?

Thanks,
Stavros

ST

Responses