This is a discussion on [rrd-users] Re: Unexpected NaN values within the RRD Users forums, part of the Networking and Network Related category; On Fri, Nov 03, 2006 at 07:05:04PM -0500, jeffrey.j.petter@verizon.com wrote: [snip story about negative ...
|
|||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
|
|||
|
On Fri, Nov 03, 2006 at 07:05:04PM -0500, jeffrey.j.petter@verizon.com wrote:
[snip story about negative numbers] (could you please quote in a proper way, so I can do too? Thanks) The problem seems to be that a number that should be treated as an unsigned integer, is treated as a signed integer somewhere in the chain. This could be your device, or it could be your monitoring program. It's like a mileage counter rolling over from 499,999 to -500,000. For a one-byte counter, the numbers would look like: .... 125, 126, 127, -128, -127, -126 ... To solve, 256 - (negative number multiplied by -1) which is: 2^8 - (negative number multiplied by -1) A two byte counter: .... 32764, 32765, 32766, 32767, -32768, -32767, -32766 ... To solve, 65536 - (negative number multiplied by -1) which is: 2^16 - (negative number multiplied by -1) One byte counters: 2^8 - (negative number multiplied by -1) Two byte counters: 2^16 - (negative number multiplied by -1) Four byte counters: 2^32 - (negative number multiplied by -1) Eight byte counters: 2^64 - (negative number multiplied by -1) Find out which one it is (it won't be one or two) and work around it in your script. Alternatively, reset the counters as soon as they are going into the danger zone. This may mean rebooting your device. I don't think rrdtool should cope with this. Would it need to assume the counter to be 1, 2, 4 or another amount of bytes? Of course, hunting down the bug and fixing it at the proper place in stead of working around it would be nice. Try fetching values from your device manually if you can. Don't use the same software your monitoring program uses right now. Still getting negative numbers? Then the device is suspect. Getting positive numbers? Then the snmp query program is suspect. HTH -- Alex van den Bogaerdt http://www.vandenbogaerdt.nl/rrdtool/ -- Unsubscribe mailto:rrd-users-request@list.ee.ethz.ch?subject=unsubscribe Help mailto:rrd-users-request@list.ee.ethz.ch?subject=help Archive http://lists.ee.ethz.ch/rrd-users WebAdmin http://lists.ee.ethz.ch/lsg2.cgi |