Michael Dingwall wrote:
Hey guys,
Thanks for the help. Found out that the owner for the rrds had been changed
to nobody. That really screwed it up. Also, I don't think that they have be
owned by the apache user, because they show up when owned by the root. So, as
I just told you the pictures are showing up but there is now no information
being seen ( no lines or area graphs) so I still need some help.
I did do the telnet into port 8649 and got an xml description of the cluster
so that is working.
Checked for a second gmetad running and there was; killed it.
Changed the owner of the rrds.
Now have got to have the information placed into the graphs.
There *should* be two gmetads running - one thread listens on 8651, the
other polls/parses your list of sources (in
$GMETAD_ROOT/etc/gmetad_sources). Kill one and the whole thing breaks.
Try to telnet to the gmetad port. You should see a whole mess o' data.
Check the filesize and modification times on the RRD files. They should be
nonzero and updating fairly frequently (once a minute, at least).
You can try using RRDtool to dump the contents of the RRDs themselves (this
will tell you if the data is being parsed and written to the RRDs
successfully).
Remember that gmetad setuid's to "nobody" when it starts up. If you need
it to be someone else, edit gmetad and change that line to whatever user
you want to run it as (as I said earlier, I recommend the web user).
And if anyone cares, *my* cause of herring-bone graphs seems to be that I
was running the CVS gmond, muted, to gather data on 2.3.x/2.4.x gmonds.
Oops. Switched to the same version as the others and everything's fine. :)
