> On Wed, Nov 16, 2011 at 8:16 AM, Adam Armstrong <
adama@memetic.org>
wrote:
>
>>
>> Same thing, I should imagine. I've never seen this bug present anywhere
>> else.
>>
>> Is the device being polled quite far away? Or attached on the end of
some
>> string?
>>
>> What is the measurement directly before the spike?
>>
>> adam.
>>
>>
> It's actually diversely routed bailing wire. :) It's about ~300 miles
away
> or 6ms as the packet flies.
>
> Looking at the data it looks like in_delta is looking suspicious, it's
> either 0, 64, 896, or some crazy large value during the spike. There is
> essentially 0 inbound traffic on this interface and barely any outbound,
> could that play into it?
>
> Here are the values before and at the start of the large blocks of high
> values:
>
> 31, 2011-11-10 16:55:07, 300, 400, 64, 336
> 31, 2011-11-10 17:00:06, 299, 309525032052119, 309525032051591, 528
>
>
> 31, 2011-11-14 07:40:09, 299, 2608, 64, 2544
> 31, 2011-11-14 07:45:11, 302, 321074668856640, 321074668848076, 8564