Fwd: [External] consult OBMC SEL log solution

Lei Yu yulei.sh at bytedance.com
Mon Jul 5 15:09:11 AEST 2021


Hi Jenny,

I am copying this email to the openbmc mailing list so the community
knows the discussion.

On Thu, Jul 1, 2021 at 4:50 PM Chen, Jenny <jenny.chen at intel.com> wrote:
>
> Hi Lei,
>
>
>
> This is Jenny from Intel BMC, I saw you have some refine for IPMI SEL log, and could you kindly help answer my questions to help me know more IPMI SEL logs? Thanks in advance!
>
>
>
> I saw you added “SEL_LOGGER_SEND_TO_LOGGING_SERVICE” to control use which solution:
>
> a. The logging entry on Dbus
>
> b. The ipmi_sel log stored in "/var/log/ipmi_sel"
>
>
>
> Questions:
>
> Currently seems defaue use the b(log files) solutions. Can I know whether there is issues found and we translate to use a solutions?
> What’s currently status of solution a implementation?

The way to store "/var/log/ipmi_sel" has some issues:
1. The upstream ipmid does not support parsing this file. The
implementation is in https://github.com/openbmc/intel-ipmi-oem.
2. If the user uses intel-ipmi-oem to support the ipmi_sel log, it
introduces many other ipmi commands that are likely not needed.
3. It separates the ipmi_sel and redfish event logs. E.g. in Intel's
implementation, an SEL sent from BIOS to BMC is stored as redfish
event log in "/var/log/redfish", and no ipmi sel is created.

The way I proposed[1] could solve the above issues:
1. The upstream ipmid already supports parsing the log entries;
2. No intel-ipmi-oem is needed.
3. The log could be represented as ipmi sel or redfish event logs
(with -Dredfish-dbus-log=enabled)

[1]: https://lore.kernel.org/openbmc/CAGm54UHdTyPh0hGiatA44whg2PdRzJFg0gBnVHA4moYRr=aT=g@mail.gmail.com/


--
BRs,
Lei YU


-- 
BRs,
Lei YU


More information about the openbmc mailing list