[collectd] Another ntpd problem

Luboš Staněk lubek at users.sourceforge.net
Tue Nov 7 22:18:45 CET 2006


Hi Florian,

Florian Forster napsal(a):
> Hi Lubos,
> 
> On Mon, Nov 06, 2006 at 10:03:56PM +0100, Lubo?? Stan??k wrote:
>> So the plugin should not try to collect server whose refid is "INIT"
>> and the rest of fields is zero(?).
> 
> I'll look into this on the weekend. If, by chance, you know how
> initializing remotes can be recognized, I'd apprechiate the hint ;)
> 

I think that refid field will contain INIT (really the text "INIT" :) )
and the reach, delay, offset and dispersion will contain 0.

The reach field is zero either from the start or it will zero bits in
sequence from all ones (377 - eight connections) to all zeros (0 - no
connection).

It is useless to collect zeros.


>> And the collectd is permanently dying on this:
>> Nov  6 21:14:05 test1 collectd[29258]: rrd_update failed:
>>   ntpd/time_offset-lx.ujf.cas.cz.rrd: illegal attempt to update using
>>   time 1162844044 when last update time is 1162844044 (minimum one
>>   second step)
> 
> I certainly hope collectd isn't _dying_ on this..? Usually this message
> is (only) a warning, not even an error..
> 

Sorry. Do not solve this. It was my mistake. I did too much changes
switching between 3.10.3 and patches.

I would like to switch to master to shut up plugins but I am not brave
enough.

Best regards,
Lubos



More information about the collectd mailing list