I've put together a set of Disk IO graphs for Windows based on
netsnmp.diskio.bytes/ops/opsize, and everything displays properly
except for the Read Size portion of opsize. This may be due to a
divide by zero problem, but I'm not familiar enough with RRD to track
it down myself.
I've built report.custom.sinf.diskio.bytes,
report.custom.sinf.diskio.ops, and report.custom.sinf.diskio.opsize as
almost exact mimics of their netsnmp counterparts. I've included all
three configurations at the bottom of this email.
diskio.ops and diskio.bytes work perfectly. diskio.opsize complains
"No Data Found." The GPRINT functions display properly for the Write
side of things, but the Read ones show:
"Avg: NaN ? Min: 312.91 Max: NaN ?"
If the Read portion is commented out, the Write-half displays
perfectly. If I change the line:
The Read portion displays a graph identical to sinf.diskio.bytes (as
expected). What I don't quite understand is why this is a problem for
this graph in particular. The identical netsnmp graphs have no problem
displaying despite large empty spots in their diskio.bytes graphs.
What am I missing, or how can I go about tracking this down?