Re: Generate Crashdump log via Redfish

["Bills, Jason M" <jason.m.bills@linux.intel.com>]

On 8/14/2020 7:02 AM, Venkata Chalapathy wrote:
> Hi Everyone,
> 
> I’m looking for guidance on generating a crashdump log via the Redfish 
> interface. There are two actions available to the user to generate the 
> dump – OnDemand and SendRawPeci. But I’m not aware of the payload data I 
> need to be sending along with the request to generate the dump in either 
> of the cases. Could you please share the payload that I need to attach 
> to the request.
> 
> "#Crashdump.OnDemand": {
> 
> "target": 
> "/redfish/v1/Systems/system/LogServices/Crashdump/Actions/Oem/Crashdump.OnDemand"
> 
>        },
This is the OEM action to trigger the Intel crashdump application. 
There is no payload required for this action.

It's on my list to move this to the new standard dump service "Create" 
action when it's finalized.

> 
> "#Crashdump.SendRawPeci": {
> 
> "target": 
> "/redfish/v1/Systems/system/LogServices/Crashdump/Actions/Oem/Crashdump.SendRawPeci"
> 
PECI is a communication protocol for BMC on Intel processors.  This is 
an OEM action to send a PECI command to the CPU.  The payload is a 
single "PECICommands" JSON object that is a double array of bytes where 
each row is a raw PECI command.
{
     "PECICommands": [[48, 5, 9, 161, 0, 0, 0, 0],
                      [48, 5, 9, 161, 0, 0, 4, 0]]
}

We had proposed a standard Processor commands resource but the DMTF 
rejected it and proposed we keep this as OEM.  We're still working out 
how to handle it, but I'd like to make this standard as well.

> Best Regards,
> 
> Venka
>