Your GRAPHS setting should not have =ncv, only TEST2RRD should. In other words, it should be like this: GRAPHS="la,disk,inode,qtree,files,processes,memory,users,vmstat,iostat,tcp.http,tcp,ncv,netstat,ifstat,mrtg::1,ports,temperature,ntpstat,apache,bind,sendmail,mailq,socks,bea,iishealth,citrix,bbgen,bbtest,bbproxy,hobbitd,clock,lines,pcpuIdle"

The rest looks fine. Of course make sure you define a [pcpuIdle] section in hobbitgraph.cfg as well. I believe you also have to restart Hobbit after any changes to hobbitserver.cfg.

-Charles

Shifter wrote:

I don't quite follow Charles. Here's the relevant lines from my hobbitserver.cfg file:

TEST2RRD="cpu=la,disk,inode,qtree,memory,$PINGCOLUMN=tcp,http=tcp,dns=tcp,dig=tcp,time=ntpstat,vmstat,iostat,netstat,temperature,apache,bind,sendmail,mailq,nmailq=mailq,socks,bea,iishealth,citrix,bbgen,bbtest,bbproxy,hobbitd,files,procs=processes,ports,clock,lines,pcpuIdle=ncv"

GRAPHS="la,disk,inode,qtree,files,processes,memory,users,vmstat,iostat,tcp.http,tcp,ncv,netstat,ifstat,mrtg::1,ports,temperature,ntpstat,apache,bind,sendmail,mailq,socks,bea,iishealth,citrix,bbgen,bbtest,bbproxy,hobbitd,clock,lines,pcpuIdle=ncv"

NCV_pcpuIdle="pcpuIdle:GAUGE"

The message I'm sending from the client is a data message, not a status message. This is what's coming from my client via an ext script running every 5 minutes:

$BB $BBDISP "data $MACHINE.pcpu cat /tmp/pcpu.txt " Sage

On 11/8/06, *Charles Goyard* <cgoyard at cvf.fr <mailto:cgoyard at cvf.fr>> wrote:

>
> On the server side of things I've modified the hobbitserver.cfg
files'
> TEST2RRD and GRAPHS lines to append "pcpu_idle=ncv".
> I also have this line in the same file:
NCV_pcpu_idle="pcpu_idle:GAUGE"

The problem comes from that your setup should read as :

TEST2RRD="$TEST2RRD,pcpu=ncv"
                    ^-- match the name of the status
NCV_pcpu="pcpu_idle:GAUGE"
    ^-- same thing here


Regards,

--

Ahhhh, that makes sense. Changing my client script to:

$BB $BBDISP "data $MACHINE.pcpuIdle cat /tmp/pcpu.txt "

caused the rrd's to be created and populated with data.

The change suggested by Charles Jones was helpful as it caused a link to appear on my trends page where there was no representation of my data at all, but not the actual graph. The link doesn't produce any graphs either, but I'm getting closer. I remember reading about that elsewhere in the forums so I'll go research that a bit. Thank you all very much for the assistance.

Sage

On 11/8/06, Charles Goyard <cgoyard at cvf.fr> wrote:

Hi,

Shifter wrote :

...

$BB $BBDISP "data $MACHINE.pcpu cat /tmp/pcpu.txt

what I'm telling is that the name after "$MACHINE" must match the name used in TESTS2RRD and NCV_foobar. I think you're mixing the name of a status (or data) report (here "pcpu") and the name of a particular line (or data source, or DS) in the body of that status/data message.

Regards,

-- Charles Goyard - cgoyard at cvf.fr - (+33) 1 45 38 01 31

To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk

Thanks for the update. It seems a good portion of the traffic on this list deals strictly with the NCV stuff. While the documentation is good, there are a lot of variables and certainly a lot of restrictions (underscores, etc) that can play a vital role in whether an NCV is working or not. Overcoming the complexity is sure a good feeling, but it's awfully frustrating when you can't get it working. Unfortunately, no amount of documentation is going to be able to overcome human errors and human emotions. Even I've spent days trying to figure out why things didn't work only to realize I was spelling ncv as nvc.

My advice for future people doing this sort of stuff is to first copy a working example, before attempting your own tanget. You can then control for your variables and experiment from there. Is the wiki for hobbit official yet? Maybe we can use that for bootstrapping our fellow hobbitiers?

-Dan

On 11/8/06, Shifter <the.shifter at gmail.com> wrote:

As a final message I was able to resolve my broken graph. The graph definition in my hobbitgraph.cfg file was listing an incorrect .rrd file. Adhereing to Mr. Goyard's earlier suggestion I changed the def line to reflect the actual rrd file name ( pcpuIdle.rrd) and *poof* graph appeared in trends. Thank you again.

Sage

On 11/8/06, Shifter <the.shifter at gmail.com > wrote:

...

Ahhhh, that makes sense. Changing my client script to:

$BB $BBDISP "data $MACHINE.pcpuIdle cat /tmp/pcpu.txt "

caused the rrd's to be created and populated with data.

The change suggested by Charles Jones was helpful as it caused a link to appear on my trends page where there was no representation of my data at all, but not the actual graph. The link doesn't produce any graphs either, but I'm getting closer. I remember reading about that elsewhere in the forums so I'll go research that a bit. Thank you all very much for the assistance.

Sage

On 11/8/06, Charles Goyard < cgoyard at cvf.fr> wrote:

...

Hi,

Shifter wrote :

...

$BB $BBDISP "data $MACHINE.pcpu cat /tmp/pcpu.txt

what I'm telling is that the name after "$MACHINE" must match the name used in TESTS2RRD and NCV_foobar. I think you're mixing the name of a status (or data) report (here "pcpu") and the name of a particular line (or data source, or DS) in the body of that status/data message.

Regards,

-- Charles Goyard - cgoyard at cvf.fr - (+33) 1 45 38 01 31

To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk