[hobbit] Sending data from script to an rrd

[Ralph Mitchell]

On 10/18/06, Rob Munsch <rmunsch at solutionsforprogress.com> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Ralph Mitchell wrote:
> > On 10/18/06, Rob Munsch <rmunsch at solutionsforprogress.com> wrote:
> >> > When you add the --extra-script option in the [rrdstatus] block,
> >>
> >> Well i'm stuck already.  What file is that block in?  I can't find it in
> >> 'server.cfg nor 'graph.cfg...
> >
> > It's in server/etc/hobbitlaunch.cfg
>
> yeah found that with a grep rrdstatus, heh oops.
>
> >> So something like
> >>
> >> echo "DS:cputemp:GAUGE:600:0:U"
> >> echo "cputemp.rrd"
> >>
> >> probably isn't far off after all?
> >
> > That should do it, yes.  That's the format for the RRD creation, and
> > the name to create.  If you just want to push out one value, you'd
> > follow that with:
> >
> >   echo "$temp"
>
> I seem to still be doing something wrong...
>
> > I don't think hobbit_rrd will recreate the rrd if it already exists,
> > so you'd have to delete it sometime to get it rebuilt with the extra
> > variables.  Wave bye-bye to your old data...
>
> "Sir, I have no data."
>
> Ok, so i did this
>
> [cputemp]
>         TITLE CPU Temp
>         YAXIS Degrees Fahrenheit
>         DEF:avg=cputemp.rrd:cputemp:AVERAGE
>         AREA:cputemp#00CC00:CPU Temp Average
>         -u 1.0
>         GPRINT:cputemp:LAST: \: %5.11f (cur)
>         GPRINT:cputemp:MAX: \: %5.1lf (max)
>         GPRINT:cputemp:MIN: \: %5.1lf (min)
>         GPRINT:cputemp:AVERAGE: \: %5.1lf (avg)\n
>
> and this
>
> GRAPHS="la,disk,inode,qtree,memory,users,cputemp,.....8<...
>
> as well, on the server.
> but what i get is a text link called
>
> Status unchanged in 1 hours, 41 minutes
> Status message received from 10.10.10.12
> hobbit graph cputemp
>
> that leads to nothing.  Not even a "nan" graph :P
>
> I think maybe by echo is in the wrong place...?  Should it be before or
> after the $BB msg command?  Does it matter?
>
> hobbit at currant:~/client/ext$ cat cputemp.sh
> #!/bin/sh
>
> umask 133
>
> COLUMN=cputemp
> COLOR=green
> MSG="Current CPU Temperature: "
> host=`hostname`
>
> ipmitool sensor | grep Temp | awk -F\| '{print $1,$2,$3,$4}' >
> $BBTMP/cputemp.$host
>
> cpu_temp=`tail -n 4 $BBTMP/cputemp.$host |head -n 1 | awk '{print $2}'
> |sed 's/\..*//g'`
> # Converting to Fahrenheit.
> temp=$((${cpu_temp}*9/5+32))
>
> # Ok, we have the temp in degrees as an integer.
> # Now to see if it's Bad, or even Very Bad.
>
> if $temp > 130
> then
>         COLOR=red
>         MSG="${MSG} $temp degrees Fahrenheit!
>
>         Hothothot!
>         "
> else
>         MSG="${MSG} $temp Degrees Fahrenheit
>
>         Chillin' like a Villain.
>         "
> fi
>
> echo "DS:cputemp:GAUGE:600:0:U"
> echo "cputemp.rrd"
> echo "$temp"
>
> # Get this to the hobbit server
> $BB $BBDISP "status $MACHINE.$COLUMN $COLOR `date`
>
> ${MSG}
> "
>
> exit 0

There should be *two* scripts...  The data collecting script that
produces the report and uses "$BB BBDISP....." to send the data to
Hobbit, and a separate script that echoes out the RRD creation bits
and pieces.  The first one would be launched by hobbit or cron or some
other means, and the second is executed by hobbit_rrd due to the
--extra-script flag in '[rrdstatus]'.  *That's* the one where the
"echo $temp" goes...

Ralph