[Orca-dev] another rrdtool/Orca question to ponder...

[Jon A. Tankersley]

I'm just about ready with the generation of alternate date strings with rrdtool, but while I
was working on a report, I figured out something else that would be 'very nice to have'.
I've been 'looking' at the graphs to get the 'min/max/avg' information for a text reports.  I
figured that I could fix up another optional output file to dump the 'LEGEND, and maybe
CONTENT' information to a text file (to keep it simple, just dumping the output to the file
after the gdImageString call is made for each legend element in rrd_graph.c).

I have it coded in rrdtool to be a compile option (configure option), but I've run into a
couple of 'stumbling' blocks...

1) I have the hooks in to make this a flag (rrdtool graph .... -m), but I couldn't get Orca
to pass it and get recognized, even though I tried coding it in ImageFile.pm appropriately.
It worked from the command line.

2) I have the hooks in rrd_graph to 'dump' the data, if and only if the image file is NOT
stdout, but I can't seem to get it to 'work' when called by Orca through the Perl hooks.  It
works fine with a command line invocation.

I've had trouble trying to debug the perl hooks and I've been unable to get rrdtool to open
some 'other' files I've tried to use to debug the problem (FILE *xx =
fopen("/tmp/ta_junk","a");  fprintf(xx, "some string\n");  kind of stuff)

Anyone have any idea where I'm missing things?

I'm also pondering hacking in a HRULE/VRULE tags for Orca plots to allow those options in RRD
to be supported rather than creating an rrd file with a constant  for thresholds (say I want
a 90% line on CPU utilization).  Any feedback from anyone?