Originally Posted by Ignition

Nah those pings are messy but they aren't extreme ;) Upstream utilisation of no more than perhaps 65-70% will cause those. Downstream would have to be pretty close to full.

The intermittent huge spikes made me think CPU though.

OP you've an engineer coming, good, not sure what they're going to be able to do but maybe they'll find something that isn't helping :)

Originally Posted by IanUK

Thats the second NTL employee to admit there is a serious problem in certain areas, shame their management can't be as honest and give us a statement about what they plan to do about it, speeds to the USA are sub 3 meg at peak times, which is worse than my old 3 meg connection, and I'm one of the lucky ones, at least my UK speeds are ok (for now)

Why are they continuing to take on new 10meg customers, isn't that illegal ?
They are selling something they cannot provide in those areas.

Sigh, I guess another Watchdog appearance beckons for NTL, they should get a season ticket for that program, their uselessness knows no bounds :)

Originally Posted by Ignition

High CPU utilisation on you uBR, the ping spikes are being caused when certain routing protocols run, causing the CPU to go up to max and causing issues with the uBR's handling of your traffic.

Solution is for ntl to upgrade the CPU.

That's my guesstimate of what's wrong anyway.

Originally Posted by sollp

I did a quick ping and seemed ok at the time.

---------- Post added at 22:42 ---------- Previous post was at 22:39 ----------



Just to put things into perpective, the service tech that visited will not know or have been told any info about these issues that are occuring. He is just giving an opinion, not based on fact or anything else.

Originally Posted by Chrysalis

Ignition when you left was their ways for people to monitor cpu utilisation no the ubr's?

Originally Posted by sollp

I did a quick ping and seemed ok at the time.

Originally Posted by Ignition

Yep.

---------- Post added at 10:08 ---------- Previous post was at 10:06 ----------



Key words there 'quick' and 'at the time' though dude :(

Not sure what you have access to though, I rarely tested by pinging in that manner as it's too open to customer traffic making it go higher (as a certain person I know used to do to avoid paying his bill), I tended to use extended DOCSIS ping or just look at minislot utilisation on the upstream. If uBR CPU was running high it was very apparent from being logged onto it.

Stats from the BADGER might be handy too if you can't test at 'peak' loading times, along with loads of other pretty stats knocking about the place.

Originally Posted by Chrysalis

Ignition when you left was their ways for people to monitor cpu utilisation no the ubr's?

---------- Post added at 06:59 ---------- Previous post was at 06:57 ----------



Sollp then what is your opinion of what would cause performance deteriation at peak hours only?