IPMI LAN command story design - try 3rd send
tomjose at linux.vnet.ibm.com
Fri Dec 15 00:03:32 AEDT 2017
On Thursday 14 December 2017 06:00 PM, Jeremy Kerr wrote:
> Hi Tom,
>>> Requiring non-standard command to be issued causes a disconnect between
>>> user expectations and our implementation. We'd have to document this
>>> difference in behaviour for every potential sysadmin interacting with
>> The command to apply BMC network settings on OpenBMC IPMI reference
>> implementation is set channel access, which is a standard command
>> mentioned in the IPMI specification.
> It may be a standard command, but using it in this context certainly
> isn't standard, nor even a common convention.
> The issue is that we're introducing an "unusual" step for this
> configuration process, and expecting all users to know about it.
>> It is not clear to me from the IPMI specification, on which
>> the BMC network settings are applied.
> Ideally, the settings should be applied as soon as the 'Set in progress'
> flag indicates that the change is complete.
> Adding the 10-second delay allows users to "batch" multiple changes,
> meaning that network connectivity isn't lost between those changes. This
> is purely to allow ipmitool (which doesn't have a facility to make
> multiple changes before indicating that the change is complete) to be
> used over the network to change the lan configuration (which is a bit
> risky regardless...).
Yeah all this lack of clarity, arises from a lack of mechanism in IPMI to
apply multiple changes (IP address, netmask gateway) together. In the
ipmitool implementation the set-in-progress is cleared after each network
> For doing this over the in-band interface, there is no harm in applying
> the settings immediately.
> To implement something complying to a "principle of least surprise", I'd
> suggest that we should:
> - batch changes, and have them applied after 10 seconds from clearing
> the set-in-progress flag
> - submit patches to ipmitool to allow multiple changes to be applied
> at once (ie, one clearing of the set-in-progress flag).
I prefer the second option of allowing ipmitool to explicitly clear the
thus allowing multiple changes to take effect.
The first option is really "unusual way" to get this done, to overcome
a specification shortcoming.
More information about the openbmc