[Orca-users] Re: plot data cut off

[Blair Zajac]

blum0422 wrote:
> 
> Blair,
>      I thank you very much for your support.  I played around with
> the summary_format and finally got it to work properly (mistake on my
> part :-( )  I did however try the plot width and that worked as well.
>
> On a side note (question #2) I'm seeing that some of the hosts create
> multiple graphs for same dataset (disk use percent).  I've now figured
> out from searching through this forum that this is due to different plots
> taken at different times.  I've now upgraded the orcallator.se on all
> clients and the orca host and removed the old rrd's and rebuilt (with
> data dating back to Jan) and still they show multiple graphs. 

Right.  This is an issue with Orca in how it matches up different sets
of data sources to put them in a single plot.  Anytime one data source
(disk in this case) is removed or added, it generates a new "plot" since
a plot is a list of its data sources.

Orca needs to be fixed to do this.

> I've read the rrd docs on 'rrdtool update' but want to know if that is the
> only way to resolve my issue while maintaining historical data.

What exactly are you trying to do here?  Use RRD to store all data?
Realize that RRD looses resolution over time as data gets older.  It
consolidates multiple data points into fewer data points at pre-defined
data ages.


> If I create the template for the hosts and (using veritas clustering
> for failover) the disks fail over to another host, will that agian alter
> the DB?

You mean the RRD DB?  Which DB?

Also, Orca uses a single RRD file for each unique measurement and combines
them when it makes the plots.

> If there is a measurement in rrd that is not included in the
> percol/orcallator file that updates the rrd will that cause conflicts? 

I'm not sure what you mean here.  If there's no data in the percol or
orcallator files, then the RRD is not updated.

> This is more or less a production tool that many people are looking at
> and don't have a lab box that I can play with :-( so would like to be
> pretty sure before I make any abrupt changes ;-) 
> Thanks again for all your support.

Best,
Blair

-- 
Blair Zajac <blair at orcaware.com>
Web and OS performance plots - http://www.orcaware.com/orca/