DNS Timeout

[Chris1949]

Hi everyone,
I live in rural France and at long last have accesss via WiFi to ADSl. Only problem it cuts out about every 30 minutes. A ping test to IP shows no problem but to the DNS primary, initially 70ms then gradually increases to 120, 200, etc until 'timeout'. It then remains offline for about 3 minutes. Our supplier seems either reluctant to address the problem or incapable. Does anyone have any ideas what might be the problem?
Thanks in anticipation,
Chris.

[JosAH]

Do you have the same problems when you're directly hooked up to that ADSL line, i.e. no WiFi involved (if that is possible for you)? If you can connect directly and don't experience those problems, blame it on your WiFi connection and try another band (there are 13 of them available in Europe).

WiFi is extremely sensitive to relative humidity and it can build up slow oscillating waves that ruin your connection. What's distance between your WiFi access point and your ADSL wired connection? Are they in the line of sight?

kind regards,

Jos

[Chris1949]

Thanks for your reply. WiFi transmitter in line of site, about 1.86 miles. Ping test on IP 10.10.10.15 no problem. Ping test on 85.31.192.22 causes problems. It doesn't appear to make any difference if it is dry or wet.
Chris.

[Ciary]

@Chris1949
i'm not sure but isnt 10.10.10.15 an inside IPaddress? and 85.31.192.22 an outside IPaddress? if thats the case, the problem will probably your router. i'm not sure if you have access to it but if you do, check NAT/PAT, DNS and physical interfaces (are the interfaces on or off and should they be what they are) after that check routingprotocol (RIP, (E)IGRP(if cisco or linksys), OSPF,...)

[Chris1949]

Hi,
Quite right 10.10.10.15 is internal. The hoster address is 10.10.10.1
Hoster is called MESHNET.fr

[Ciary]

then its probably a routing problem. do you have access to the router?

[NeoPa]

I suggest it would be useful if you describe the topology (what links to where) of your network, with IP addresses where possible. This will give us a better chance of understanding what may be happening in the system and/or devising tests that may help determine where the issue lies.

[NeoPa]

Sometimes PINGs can be unreliable. Various routing devices spoof addresses they think they know they deal with. Essentially they say "I've seen this work before, so rather than waste the traffic shooting it off I'll reply more quickly myself, pretending to be the device requested."

I suggest PathPing or TraceRt instead.

[JosAH]

Is there someone else in the vicinity using the same (or near) band on the Wifi band (+- 4GHz)?

Maybe there's an old micro wave oven nearby ...

Another approach could be to install larger antennas on both sides; maybe you can use a couple of extra dBs.

kind regards,

Jos

[NeoPa]

Without further detailed information I would think Jos is most likely on the right lines here.
@JosAH
Don't let him confuse you with reference to databases. He's talking about Decibels :D

[JosAH]

@NeoPa
I'm always on the right line unless I'm off line occasionally ;-) The OP mentioned 1.86 miles which is quite a large distance. Normal antennas only operate well within a distance of 80 ft or so (in the line of sight). I'm tempted to suggest a parabolic dish antenna or even a Yagis antenna. Those little stick antennas just don't make it over distances like that.

BTW I was talking about 4GHz, that should be 2.4GHz of course.

kind regards,

Jos

[Chris1949]

Hi,
Thanks for all your efforts, here's the technical info direct from the 'meshnet' site.
Device Type : 5.4GHz - Subscriber Module - 0a-00-3e-54-09-6b
Software Version : CANOPY 8.2.7 SM-DES
Software BOOT Version : CANOPYBOOT 1.0
Board Type : P10
FPGA Version : 061708
PLD Version : 11
Uptime : 6d, 01:06:04
System Time : 01:06:02 01/07/2001
Ethernet Interface : 100Base-TX Full Duplex
Antenna : Vertical
Subscriber Module Stats
Session Status : REGISTERED VC 19 Rate 1X/1X
Registered AP : 0a-00-3e-54-04-7b
RSSI : 555
Power Level : -77 dBm
Jitter : 3
Air Delay : 182 approximately 1.69 miles (8918 feet)

[NeoPa]

@NeoPa
That may prove to be helpful info, but I really need answers to my questions in post #7 (quoted) to be able to consider the problem clearly.

[Chris1949]

Hi,
I have an office pc and a laptop connected to a d-link ethernet switch. In turn this is directly connected to the antenna fixed on the wall and pointed directly at the antenna at 1.69 miles. From there the signal is transmitted about 4.5 miles to a large radio transmitter and the signal forwarded on to the town of Aurillac. Other users go via a different antenna to the radio transmitter and don't appear to have the same problems. It would appear that the problem lies within the control system associated with our local antenna that serves about 20 customers.
Hope this info helps you,
Chris.

[NeoPa]

So none of your equipment uses wireless technology to connect with your local D-Link ethernet switch?

Please also provide any IP addresses that you have. Internal & external for the switch if possible.

[Chris1949]

No, both are by cable. PC I.P. 10.10.10.15 Laptop I.P. 10.10.10.13
Default passerelle 10.10.10.1 Staic I.P. used as Meshnet does not accept dynamic I.P.s

[NeoPa]

It seems like the only WiFi involved is your link to your provider, via equipment entirely configured by them. Everything on your internal network is wired.

I see nothing here that could be the problem other than the service provided for you. I would think they would need to run some diagnostics themselves to discover the issue if they were interested enough.

You could try using TraceRt to your DNS server when it's running ok, and again when it's not, to compare the two. Unfortunately I doubt that would help much until you could convince your proviser to look into it for you.

[Chris1949]

Thanks for your efforts. I've tried TRACEROUTE from numion .com and not once has it managed to complete the trace. It always stops one step from home. Meshnet have said they are looking at the EDF electricity supply! I cannot imagine how any fluctuations can have the observed effect. We'll wait and see.
Chris.

[NeoPa]

Sometimes the rate at which the electricity pulses (50 Hz etc) can have an adverse effect on wave energy transmissions (WiFi). Alternatively the reliability of the supply may be an issue. I'm guessing of course, but there are often technical issues we wouldn't anticipate unless we had the relevant experience.

[Chris1949]

Hi,
Apparently they are going tocheck the supply this week. Personally I can't imagine that even if it is pulsing it would create the observed problems. I'll keep you informed.
Have a nice day,
Chris.

[NeoPa]

Hi Chris. Good to hear they're looking at it at least.

I didn't mean to suggest that was necessarily the cause. Just illustrating there are often effects we uninitiated would be unaware of. From the frequencies involved I very much doubt it, but this is an area I'd certainly leave to the experts to sort out (if they will).

[Chris1949]

Hi,
One strange aspect of the problem is that it seems to go off nearly always 'on the hour' a few minutes either way. Off for 3 minutes and then back on. Bizarre.
Chris.

[NeoPa]

I have to say that sounds more and more like something, somewhere, is interfering with something. Probably signal related.

Could be in the local environment. Equally could be something service related.

Let's hope the provider puts us out of our curiosity misery eh.

[Chris1949]

Just don't hold your breath waiting. The installation was 18 months late!!

[NeoPa]

Ah :( Not so promising then.

Good luck anyway, and if they do ever resolve it I expect we're all curious to know what it eventually turns out to be.

[JosAH]

@Chris1949
That's fast enough ;-) so (according to my calculations) somewhere around 2012 all your problems are solved! Too bad the universe collapses by that day (or so the told me ;-)

kind regards,

Jos

[Chris1949]

A quick update on my MESHNET connection. There is no update. No improvement, no information, no good.
Will keep you posted.
Chris.

[NeoPa]

Sorry to hear it Chris :(

I moved when my ISP was that useless. If you have the option I'd recommend it.

[Chris1949]

Hi,
No other option but to hope they will improve the service. Download speed last night fell to 44k!!! Its supposed to be 1M.
Perhaps when they realise that most clients have stopped the Direct Debits then they may take the situation more seriously.
Have a nice weekend,
Chris.

[NeoPa]

You too Chris.

Actually, your throughput can be affected by either of two factors (purely for info) :

1. The link speed to your node.
2. The traffic going through the ISP's gateways / routers.

Typically, if it is variable, it is likely to be the latter. This may be frustrating but a claim of 1Mbps download speed doesn't mean you will necessarily achieve that level of throughput. Merely that your link will not limit it to slower than 1Mbps.

NB. Mbps is Mega-Bits Per Second. Not Mega-Bytes. This effectively relates to 1/10 MBps (Mega-Bytes Per Second). Most links are rated in Mbps.