SUMMARY: interesting hadf error message

From: Rae, Scott (
Date: Mon Apr 24 2000 - 11:25:35 CDT

This appears to be a known problem to Sun - it is bug ID 4029779

Details: Solstice HA attempts to do some primitive monitoring
> of how loaded the two hosts are. The motivation for that
> monitoring is to warn the administrator (via syslog) when the
> total load on the two hosts is so high that it is unlikely
> that after a failure the surviving host would be able to handle
> all of the load itself.
> The messages you are seeing are telling you that the rpc.rstatd
> daemon has a bug that it cannot compute the boottime dependably.
> And that due to this bug, the load monitoring is not being done.
> The load monitoring is pretty primitive anyway, so whether it is
> of any value in the first place is debateable.

I have searched SunSolve and can't find a patch for this. The message has
not appeared for a few days now since the machine was rebooted for another
reason (it was a daily occurrence before).

Many thanks to Jarrett Carver.


> -----Original Message-----
> From: Rae, Scott
> Sent: 19 April 2000 16:21
> To: 'sun managers question'
> Subject: interesting hadf error message
> I couldn't resist posting this.....
> Apr 17 14:35:32 phys-ki-sun4k1 hadf: ERROR: Boottime for host
> 'phys-ki-sun4k1' has changed, prevReading=954831583 curReading=954831538
> Apr 17 14:35:32 phys-ki-sun4k1 hadf: ERROR: Warning: fdl_load: boottime on
> host 'phys-ki-sun4k1', as reported by rpc.rstatd, has
> Apr 17 14:35:32 phys-ki-sun4k1 hadf: ERROR: moved backwards in time (this
> is a due to a design flaw in rpc.rstatd).
> Apr 17 14:35:32 phys-ki-sun4k1 hadf: ERROR: Therefore, not computing
> overload stats, exiting.
> Has anyone an idea of what might have happened? The HA partner node for
> this machine has logged the same message....
> Regards
> Scott Rae

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:14:06 CDT