[rrd-developers] XPORT command, rrd_xport.c has version problem
Barker, David
dbarker at gnilink.net
Mon Jan 19 16:45:31 MET 2004
XPORT Bug and Fix
I have run into an issue with XPORT when the requested data results in
multiple variables with different granularity (step sizes.) For example,
you might have a variable, say ifInOctets measured every 60 seconds, and a
consolidation function for Max every 5 minutes, or 300 seconds. Then you
request the latest two hours of data, you will get 121 ifInOctet values, but
only 25 Max consolidations of that variable. This should not be a problem,
but XPORT matches the first 25 Max variables with the first 25 ifInOctet
variables.
To make matters worse, if ifInOctets is listed first in the request, the Max
data for rows 26 - 121 will be invalid, or worse, XPORT will fail with a
segmentation fault.
XPORT Version Problem
I have coded a fix for this using rrdtool version 1.0.46 as a baseline. I
tried to get the development version, but the comment in the file
rrd_xport.c mentioned version 1.0.37, which didn't seem right. Here is the
current state of the code as I can see it:
Version 1.0.46 Development Version
Stamped 1.0.46 Stamped 1.0.37
Uses str2rpn() Uses rpn_parse()
numRows=(end-start)/step + 1 numRows=(end-start)/step
Improves unknown option handling Removes some 'dead' code
Uses rrd_graph_init()
I have created a patch for 1.0.46, but I cannot compile the development
version, so I'm not sure what to do. Can someone who has been contributing
to rrdtool give me some advice here? The changes are not that complicated,
but I want to make sure that I don't fix the wrong baseline (so my fixes get
lost) or stomp on someone else's fixes. Thanks.
-David Barker
Network Computing Technologies, Inc.
DBarker (at) gnilink (dot) com
703-295-4978
--
Unsubscribe mailto:rrd-developers-request at list.ee.ethz.ch?subject=unsubscribe
Help mailto:rrd-developers-request at list.ee.ethz.ch?subject=help
Archive http://www.ee.ethz.ch/~slist/rrd-developers
WebAdmin http://www.ee.ethz.ch/~slist/lsg2.cgi
More information about the rrd-developers
mailing list