perfsonar development work

[Ilias Tsompanidis <>]

Szymon Trocha wrote:
> Loukik Kudarimoti wrote:
> > Roman Lapacz wrote:
> > > I think that if rrd has been created then data were accepted. I'm 
> > > not sure about rrdtool itself how/if this keeps only single value.
>
> What's the problem?
I've been populating the rrd-ma via the Store functionality and I  ... 
can't find the data I stored.
Just for testing purposes, i let a script send Storerequests with the 
same data, for "current" timestamps , every 3 minutes all night. Then I 
passed the key to a SetupDataRequest, and all I've got was "0"... Of 
course the same happens when quering the rrd file itself via rrdtool. ( 
In a similar test case for Errors, the result was "NaN").
I've noticed that the "last update" changes, so rrd-ma updates the file. 
I don't know what exactly it is supposed to save...
In normal circumstances, rrd returns the five minute average of what has 
been stored. So storing every 3 minutes the same value should be enough 
to store that data in rrdma and be able to get it back. I would be happy 
with ANY data actually, instead of 0 or NaN.
If I'm doing something wrong, or the service wasn't able to understand 
what i sent, I think that there should be an error somewhere.
But the service seems to get to the point where it should update the rrd 
file. So I'm naturally confused...

Ilias
> > > My general opinion about rrd store functionality is that we could 
> > > remove it from the service. I don't see any interest from users to 
> > > have it because normally tools like mrtg or cricket do it in natural 
> > > way. I think fetching should be enough.
> > Perhaps you can leave the code in but move the store interface from 
> > the documented interfaces list to the undocumented ones? If and when 
> > we resurrect the SNMP polling and data storing product, this 
> > interface can be enhanced and documented as well..
>
> Which SNMP product are you talking about? I though you abandoned your 
> SNMP service...
>
> Regards,