>> >constantly updating their "2.10 Beta" code for over a year without
>> >changing the revision number. The only way to tell that there is a
>> >newer version is the date of the file itself. I used to be using
>> >09/07/06 but I've been recently using 12/27/06 and I understand there is
>> >now a 02/09/07 version. No versions really give an errata sheet to say
>> >what's changed.
> >Would it make sense to complain to SuperMicro?
I recently escalate this issue both for IPMI and their BIOS files.
According to them, the files between these dates should be identical as
the newer dates are only due to adding additional motherboard support
and the way they do releases, all the files are replaced. Yet I bet if
I ran an MD5/SHA1SUM between these different dated files, they would be
different. I'm going to try it and see what happens but it seems that
Supermicro isn't as open to errata reporting as I would like. The other
thing Supermicro said was that they only issue errata documentation once
the firmware is released (not beta) but I took exception to this because
the 2.10Beta has been in "beta" for well over 12 months without any
release. That and it's the only version with full IPMI2.0 support.
> >Broadcom, dedicated MAC shared interface.
That's a good thing. Intel makes great NICs but sharing one MAC for
multiple devices has lots of issues. Intel has since changed their
designs to use two MACs for IPMI-enabled machines:
Anyway, as a test, you might try using Supermicro's IPMI tools to see if
you see the same level of fallout compared to IPMITool:
I think the Supermicro solution is more tolerant of IPMI packet
corruption, loss, etc. compared to IPMITool. In the future, I hope we
can get the ipmitool more tolerant as well.