freeipmi-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Freeipmi-devel] Re: SEL datestamp 01-Jan-1970?


From: Al Chu
Subject: RE: [Freeipmi-devel] Re: SEL datestamp 01-Jan-1970?
Date: Wed, 18 Feb 2009 09:30:31 -0800

Sounds good to me.  Perhaps what's best is I just add a note into the
ipmi-sel manpage describing this.

Al

On Wed, 2009-02-18 at 05:53 -0800, Andy Cress wrote:
> Al,
> 
> You could use pre-init or uninitialized, both are fine and mean the
> same.
> The IPMI spec doesn't say either way.
> 
> Rather than mask out the 1-Jan-1970, however, in ipmiutil I display it
> as is, for this reason:
> If the power is removed, then restored, the BMC starts up with
> 1-Jan-1970 00:00:00.
> However, if the system is plugged in but not powered up to BIOS POST for
> a while,
> this would show up in the SEL timestamp as something like 1-Jan-1970
> 01:20:00,
> meaning that the system was plugged in an hour and 20 minutes ago, but
> not powered up yet.
> 
> Masking the real data with a "pre-init" string, or masking the actual
> sensor reading value as "unknown" is one of the faults of ipmitool, IMO.
> 
> Andy
> 
> 
> -----Original Message-----
> From: Al Chu [mailto:address@hidden 
> Sent: Tuesday, February 17, 2009 4:44 PM
> To: Andy Cress
> Cc: address@hidden; address@hidden; address@hidden
> Subject: RE: [Freeipmi-devel] Re: SEL datestamp 01-Jan-1970?
> 
> Hey Andy,
> 
> Do you know if the phrase "Pre-init timestamp" is defined by some spec
> (maybe outside of the IPMI one)?  I remember seeing that ipmitool
> outputs this when the timestamp is 0, but is this really the right
> phrase or just what an ipmitool coder picked?  I do agree that perhaps a
> "Uninitialized" or atleast something else should be output instead of
> 01-Jan-1970.
> 
> Al
> 
> On Tue, 2009-02-17 at 12:50 -0800, Andy Cress wrote:
> > Won,
> > 
> > That date is actually a 0 value, meaning 0 seconds since 1-Jan-1970.  
> > Most of the time, the BMC stays running on 5V standby power, but if
> the
> > BMC firmware restarts (e.g. after input power is lost or unplugged),
> it
> > starts over and starts its clock value at 0 until BIOS comes up and
> > gives it the real clock value.  
> > So, you'd expect to see a "Pre-init timestamp" in this case because
> the
> > SEL shows that the input power was lost.  
> > 
> > Andy
> > 
> > -----Original Message-----
> > From: address@hidden
> > [mailto:address@hidden
> On
> > Behalf Of Al Chu
> > Sent: Tuesday, February 17, 2009 12:34 PM
> > To: address@hidden
> > Cc: address@hidden; address@hidden
> > Subject: [Freeipmi-devel] Re: SEL datestamp 01-Jan-1970?
> > 
> > Hey Won,
> > 
> > Well, I don't know if "normal" is the right phrase to use :-)  From my
> > experiences, many motherboards do report incorrect timestamps.
> FreeIPMI
> > simply parses whatever information is on the motherboard.  So if it is
> > given a bad timestamp, it will output whatever that timestamp equates
> > to.
> > 
> > Al
> > 
> > On Tue, 2009-02-17 at 08:29 -0800, Won De Erick wrote:
> > > Hi All,
> > > 
> > > it seems that the CMOS battery is just fine. I've replicated this to
> > several similar boxes.
> > > 
> > > Is this related to how freeipmi parses logs? or this is just normal?
> > > 
> > > Thanks in advance.
> > > 
> > > Won
> > > 
> > > 
> > > --- On Fri, 2/13/09, Won De Erick <address@hidden> wrote:
> > > > Hi Al,
> > > > 
> > > > I noticed the following erroneous datestamps when viewing
> > > > SEL.
> > > > 
> > > > #ipmi-sel
> > > > 144:08-Feb-2009 03:51:08:System Event System
> > > > Event:Timestamp Clock Synch
> > > > 164:08-Feb-2009 03:51:03:System Event System
> > > > Event:Timestamp Clock Synch
> > > > 184:08-Feb-2009 03:51:25:System Event System Event:OEM
> > > > System Boot Event
> > > > ...
> > > > 444:01-Jan-1970 08:00:11:Power Supply Power Supply 2:Power
> > > > Supply input lost (AC/DC)
> > > > 464:01-Jan-1970 08:00:12:Power Unit Power
> > > > Redundancy:Entered from Non-redundant:Insufficient Resources
> > > > 484:01-Jan-1970 08:00:50:System Event System
> > > > Event:Timestamp Clock Synch
> > > > 504:10-Feb-2009 11:45:09:System Event System
> > > > Event:Timestamp Clock Synch
> > > > 
> > > > I've replicated this to several machines. Though the
> > > > CMOS battery could be the potential reason, I want to know
> > > > if ipmi-sel is related to this, or has something to do with
> > > > this.
> > > > 
> > > > Thanks,
> > > > 
> > > > Won
> > > 
> > > 
> > >       
> > > 
> -- 
> Albert Chu
> address@hidden
> Computer Scientist
> High Performance Systems Division
> Lawrence Livermore National Laboratory
> 
> 
> The information contained in this document is CONFIDENTIAL and property of 
> Kontron. Any unauthorized review, use, disclosure or distribution is 
> prohibited without express written consent of Kontron. If you are not the 
> intended recipient, please contact the sender and destroy all copies of the 
> original message and enclosed attachments.
> 
-- 
Albert Chu
address@hidden
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory





reply via email to

[Prev in Thread] Current Thread [Next in Thread]