[mythtv-users] how to troubleshoot network congestion problems with mythtv [solved]
linux at thehobsons.co.uk
Wed Jun 16 17:53:46 UTC 2010
Tim Coote wrote:
>For completeness, I thought that I'd note that I've fixed this:
>- the behaviour was anomalous, ping times are now stable at ~4.5ms
>and I know what to look for
>- the root cause was a Nintendo DS plugged into a mains socket. I
>don't know whether the transformer is broken or just letting through
>spikes from the DS onto the mains.
>I'm still getting odd spikes, but at least I know that this
>shouldn't happen and I ought to be able to isolate offending items.
Like I said, what you are (or were) seeing is quite likely the result
of error correction at the physical layer - and the result of having
something noisy on the mains is that the units have to employ error
correction more often. I don't know what forms of error correction
these systems use, but I suspect they have some form of forward error
correction (ie additional data added to the transmission) that will
allow for small errors to be corrected by the receiver, but beyond
that they will certainly have some form of retransmission based error
correction. All this will be done at a low level and should be
transparent to the devices trying to use the link.
But what will be happening is that every time a packet cannot be
fully decoded and corrected, then a retransmission will be needed -
and that will delay the packet(s) affected, giving you an abnormally
long round trip (ping) time.
As Brian Wood says, there are many cheap PSUs on the market that have
"lost" noise suppression components between design and manufacturer -
and they often have a board with space for the components, but none
fitted. People know that unless there is a complaint, then they can
pretty well sell what they like and get away with it - such is the
state of our standards enforcement these days where it's mostly down
to the vendor to self certify, as if that's never going to go wrong
:-/ Just look at the situation with the RSGB and OfCom - we have a
group of well qualified specialist, who have performed tests and
proved that the equipment doesn't meet standards, and OfCom turns
round and suggests that a) it's not a problem as few have complained,
and b) it's actually a result of incorrect installation (the latter
bit I find quite ... well words fail me !) So it's clear, as long as
you don't annoy too many people (or the wrong sort of well connected
people) then our regulator will ignore any wrongdoing.
As Travis Tabbal says, most people would not recognise such
interference and so are unlikely to complain about it.
Travis Tabbal wrote:
>>Quite true, and made worse by the fact that to most consumers it
>>appears that there are 10 or more channels available, when they
>>actually overlap, causing there to be just 3 actually individual
>>channels, but since people don't know that, they may well pick a
>>channel they *think* is apart from their neighbors, when in fact it
>I forgot about that. I really wish they hadn't done that with the
My memory doesn't really go that far back in technical terms, but
didn't earlier wireless standards actually stick within single
channels ? So in fact, originally there were 11 (or 13 depending on
where you live) channels available, and it would have caused lots of
confusion to try and renumber them. On the other hand, it wouldn't
have hurt if equipment manufacturers only presented the sensible
>WiFi should only be used for portable roaming machines, not
>permanent or semi-permanent installations, but again you can't
>expect the people selling such gear to tell people that.
Several times at work I've had comments from customers along the
lines of "we don't need much cable, people will be using wireless"
and I've had to tell them effectively what you wrote above - that
wireless is convenient but it's neither as fast or reliable as cable.
Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed
author Gladys Hobson. Novels - poetry - short stories - ideal as
Christmas stocking fillers. Some available as e-books.
More information about the mythtv-users