ian__b wrote: I now appear to be recording pppoa0 traffic.
I shall see what happens
Looking back over my previous results it appeared that I was seeing the problems during periods of high traffic. I therefore started up a number of video streams that, more or less, saturated my broadband connection and looked at the subsequent results from both my scanner and PRTG.
My scanner, that takes a sample via telnet every 5 minutes, saw a couple of large anomalies reported by the router; 700MB of download traffic in 5 minutes for example against the "normal" rate of around 32MB.
PRTG, getting data via SNMP every minute, seemed to get very confused; a number of readings of 4GB in a minute and 1 of 8GB. However, both of these were around significant binary numbers (0x400000 and 0x800000) so that might just be down to PRTG mishandling the values?.
Short extract of download traffic in KB (time period, my scanner, PRTG)
Code: Select all
15:00-15:05 38,373 38,458
15:05-15:10 32,336 31,989
15:10-15:15 32,243 32,522
15:15-15:20 32,271 32,640
15:20-15:25 32,557 4,226,601 *
15:25-15:30 32,772 4,226,793 *
15:30-15:35 36,309 35,758
15:35-15:40 34,534 35,175
15:40-15:45 31,325 4,225,768 *
15:45-15:50 45,399 45,545
15:50-15:55 33,396 1,235,109 *
15:55:16:00 39,925 3,032,363 *
16:00-16:05 33,050 33,342
16:05-16:10 725,641 725,494
16:10-16:15 33,174 33,153
16:15-16:20 213,071 212,299
16:20-16:25 35 8,388,631 *
Other than the starred readings, where PRTG seems to be getting confused at binary boundaries, the readings are in reasonable agreement (taking into account the slightly different time scan periods). I would guess that this means that both my telnet scan and the SNMP clients are getting the same values from the router software, and that it is the router that is generating incorrect readings.
But I could be wrong
Ian