1:57 p.m.
Recently we have begun to have problems with our browsers stopping to
work, ask it to go to a bookmark address and nothing happens, everything
stops until an error message comes up that the address can not be found.
Thunderbird continues to receive e-mail unless I restart Thunderbird,
then it is unable to find gmail.com. The problem exists on two F-19
computers and several Apple computers on the LAN, in fact I often become
aware of the problem when someone complains that the internet connection
is down.
I can always restore normal operation by re-starting the Viasat modem, a
process that takes several minutes and is an annoyance.
I have determined that I can enter a numeric address e.g. 199.106.52.212
for viasat.com and that works as expected. At the request of viasat tech
support I tried taking the router out of the circuit and connecting
directly to the modem, that effected no change. Curiously, at the time
this this morning when the problem was manifest tech support's "computer
was down" which left me wondering if there was a correlation with my
apparent DNS problem. I was asked to call back in a few hours, I did that.
The tech support person seemed to take offense at my claiming there was
a DNS problem and went on to explain that I probably have a dhcp problem
and we need to start by "re-setting my NIC." From there on the
conversation began to deteriorate and she accused me of shouting at her,
I was able to extricate myself gracefully after that and agree that I
would call back when I had more time to devote to troubleshooting.
Am I missing some point here? Is there some connection between dhcp and
dns that I am not aware of. My dhcp server is in the router and deals
with about thirty addresses on the LAN and works faultlessly as far as I
can see.
Any thoughts appreciated,
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
11:42 a.m.
On 11/12/2013 10:13 PM, Tim issued this missive:
Bob Goodwin:
> I changed the router configuration to use 208.67.229.220:5353 and
> 208.67.222.222:5353 and see the following:
>
> [root@box10 bobg]# dig www.bbc.co.uk
>
> ; <<>> DiG 9.9.3-rl.13207.22-P2-RedHat-9.9.3-5.P2.fc19 <<>>
www.bbc.co.uk
> ;; global options: +cmd
> ;; Got answer:
> ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 35870
> ;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 1
>
> ;; OPT PSEUDOSECTION:
> ; EDNS: version: 0, flags:; udp: 4096
> ;; QUESTION SECTION:
> ;www.bbc.co.uk. IN A
>
> ;; ANSWER SECTION:
> www.bbc.co.uk. 176 IN CNAME www.bbc.net.uk.
> www.bbc.net.uk. 182 IN A 212.58.246.93
> www.bbc.net.uk. 182 IN A 212.58.246.92
>
> ;; Query time: 616 msec
> ;; SERVER: 192.168.1.1#53(192.168.1.1)
> ;; WHEN: Tue Nov 12 11:46:20 EST 2013
> ;; MSG SIZE rcvd: 100
>
> As you say the cache is much faster, 2 ms as opposed to 616 ms.
>
> But I still don't know what DNS address it actually goes to, how do I
> find that?
Just to be clear, DNS means domain name *system* not *server*. So, are
you asking you're not sure what DNS server has responded, or you don't
know how to read the answers you got back from dig?
Looking at your above dig results:
Your command line asked whatever your computer's default DNS server was
(configured in /etc/resolv.conf) to look up www.bbc.co.uk.
The ANSWER section gives results about your query addresses. A query
for www.bbc.co.uk gets a CNAME answer of www.bbc.net.uk (i.e. it says
use that second name, instead). Followed by two answers for IPs for
www.bbc.net.uk, saying to use either 212.58.246.93 or 212.58.246.92
(either will do).
The QUERY TIME section says the server that answered was 192.168.1.1, so
that would be the first DNS server address in your /etc/resolv.conf
file. Second or third server addresses in that config file will only be
consulted if the first one failed to respond, and there'd be a seriously
long time-out period before that happened.
Your router, at 192.168.1.1 may have tried to query 208.67.229.220:5353
and 208.67.222.222:5353 to find the results (I don't know if your
router/modem can accept port numbers attached to the IPs, like that),
and may have connected to them. Or, your ISP may have intercepted that.
You might try using those IPs and ports with the dig command, directly.
By default, dig will use your pre-configured DNS servers (in
your /etc/resolv.conf file) unless you specify a server to query on its
command line. That is *which* server to ask, not just what address you
want to look up.
If you want to use dig to query a DNS server that is NOT in your
/etc/resolv.conf file, use something like:
dig @208.67.229.220 -p 5353 www.bbc.co.uk
The "@ip-address" would be the address of the DNS server you want to
query. If you don't want to use the standard DNS port of 53, then
include the "-p <val>" with "<val>" the non-standard port
number to use
(e.g. port 5353 in this example). See "man dig" for other options.
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer, AllDigital ricks(a)alldigital.com -
- AIM/Skype: therps2 ICQ: 22643734 Yahoo: origrps2 -
- -
- Diplomacy: The art of saying "Nice doggy!" until you can find a -
- big enough rock. -
----------------------------------------------------------------------
6:12 p.m.
On 11/12/2013 04:03 PM, Cameron Simpson wrote:
And the front line phone staff may be unaware of the issue.
Speaking as somebody who did phone tech support for an ISP, that's quite
possible. Granted, we did have a page that (supposedly) listed all
known outages, but it was almost never up to date and usually just
claimed that all was OK even when our email servers had been down for
hours, or something like that.
12:13 a.m.
Bob Goodwin:
I changed the router configuration to use 208.67.229.220:5353 and
208.67.222.222:5353 and see the following:
[root@box10 bobg]# dig www.bbc.co.uk
; <<>> DiG 9.9.3-rl.13207.22-P2-RedHat-9.9.3-5.P2.fc19 <<>>
www.bbc.co.uk
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 35870
;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;www.bbc.co.uk. IN A
;; ANSWER SECTION:
www.bbc.co.uk. 176 IN CNAME www.bbc.net.uk.
www.bbc.net.uk. 182 IN A 212.58.246.93
www.bbc.net.uk. 182 IN A 212.58.246.92
;; Query time: 616 msec
;; SERVER: 192.168.1.1#53(192.168.1.1)
;; WHEN: Tue Nov 12 11:46:20 EST 2013
;; MSG SIZE rcvd: 100
As you say the cache is much faster, 2 ms as opposed to 616 ms.
But I still don't know what DNS address it actually goes to, how do I
find that?
Just to be clear, DNS means domain name *system* not *server*. So, are
you asking you're not sure what DNS server has responded, or you don't
know how to read the answers you got back from dig?
Looking at your above dig results:
Your command line asked whatever your computer's default DNS server was
(configured in /etc/resolv.conf) to look up www.bbc.co.uk.
The ANSWER section gives results about your query addresses. A query
for www.bbc.co.uk gets a CNAME answer of www.bbc.net.uk (i.e. it says
use that second name, instead). Followed by two answers for IPs for
www.bbc.net.uk, saying to use either 212.58.246.93 or 212.58.246.92
(either will do).
The QUERY TIME section says the server that answered was 192.168.1.1, so
that would be the first DNS server address in your /etc/resolv.conf
file. Second or third server addresses in that config file will only be
consulted if the first one failed to respond, and there'd be a seriously
long time-out period before that happened.
Your router, at 192.168.1.1 may have tried to query 208.67.229.220:5353
and 208.67.222.222:5353 to find the results (I don't know if your
router/modem can accept port numbers attached to the IPs, like that),
and may have connected to them. Or, your ISP may have intercepted that.
You might try using those IPs and ports with the dig command, directly.
By default, dig will use your pre-configured DNS servers (in
your /etc/resolv.conf file) unless you specify a server to query on its
command line. That is *which* server to ask, not just what address you
want to look up.
--
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686
Don't send private replies to my address, the mailbox is ignored. I
read messages from the public lists.
6:03 p.m.
On 12Nov2013 13:19, Bob Goodwin <bobgoodwin(a)wildblue.net> wrote:
[...] I have had some intermittent problems recently but since
my complaint yesterday it has worked well. Coincidence perhaps?
Maybe, maybe not. If one of their DNS servers was flakey, your
complaint would not be alone. If they replaced/repaired their server,
everyone will see the fix.
And the front line phone staff may be unaware of the issue.
--
Cameron Simpson <cs(a)zip.com.au>
I'm a man of my word. In the end, that's all there is. - Avon
6:47 p.m.
On 12/11/13 18:58, Cameron Simpson wrote:
On 10Nov2013 20:18, Bob Goodwin <bobgoodwin(a)wildblue.net>
wrote:
> [...] Further more their system
> does other things that I don't want to lose, namely the "free time"
> during the wee hours when I can allow the Apple computers to connect
> with iCloud, [...]
Regarding reigning in the Apple stuff, what approach are you using here?
Cheers,
Most of the time I only allow connection to the iCloud, iTunes, etc.
between midnight and five local time which is what they allow without
tracking usage. I can open that if there is a need to connect at other
times.
I also have another filter that controls specific LAN addresses from
connecting to the WAN except between midnight and five.
And another that I hope blocks WAN traffic to my NFS also on the LAN.
I use the Tomato 1.28 version of DD-WRT which permits a lot of control,
logging, and I can view usage at any time even in real time which is a
good way to see what my system download/upload speeds are as stuff is
transferred.
I also keep a daily record of the ISP'slogged data for my daily usage so
I always know where I stand. I get 25 gigs per month but it costs me $10
a gig for anything I want beyond that. I've been managing our usage for
about seven years now and it works well for me.
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
12:28 p.m.
On 11/12/2013 01:19 PM, Bob Goodwin ~ Zuni, Virginia, USA wrote:
On 12/11/13 13:03, Jim wrote:
> The DNS that is in your router is what IP it goes to, which in most
> cases your Internet Provider .
Yes, normally you would expect that but Viasat Exede forces you to
their DNS no matter what you set your router for and there is nothing
wrong with the way that works as long as it is working normally. I
have had some intermittent problems recently but since my complaint
yesterday it has worked well. Coincidence perhaps?
Bob
The DNS numbers you have in /etc/resolv.conf would be the DNS-IP that it
would goto.
At least that is my understanding.
12:19 p.m.
On 12/11/13 13:03, Jim wrote:
The DNS that is in your router is what IP it goes to, which in most
cases your Internet Provider .
Yes, normally you would expect that but Viasat Exede forces you to their
DNS no matter what you set your router for and there is nothing wrong
with the way that works as long as it is working normally. I have had
some intermittent problems recently but since my complaint yesterday it
has worked well. Coincidence perhaps?
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
12:03 p.m.
On 11/12/2013 12:00 PM, Bob Goodwin ~ Zuni, Virginia, USA wrote:
On 12/11/13 09:41, Tim wrote:
> I get about a quarter or third of that time if I try other DNS servers.
> My own took 1500 mS the first time, then 6 ms for subsequent attempts
> (local cache is much quicker than internet propagation delays). Queries
> for other addresses took about 100 mS, so the BBC query must have been
> answered by something further away. My server checks it cache, then (if
> it doesn't have an answer) goes out to the root servers (to find out who
> would know the answer), then whatever servers the roots tell it, like a
> proper traditional DNS server.
>
> I consider 1.5 secs for a response to be unusually long. I couldn't put
> up with that, or worse, consistently.
I changed the router configuration to use 208.67.229.220:5353 and
208.67.222.222:5353 and see the following:
[root@box10 bobg]# dig www.bbc.co.uk
; <<>> DiG 9.9.3-rl.13207.22-P2-RedHat-9.9.3-5.P2.fc19 <<>>
www.bbc.co.uk
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 35870
;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;www.bbc.co.uk. IN A
;; ANSWER SECTION:
www.bbc.co.uk. 176 IN CNAME www.bbc.net.uk.
www.bbc.net.uk. 182 IN A 212.58.246.93
www.bbc.net.uk. 182 IN A 212.58.246.92
;; Query time: 616 msec
;; SERVER: 192.168.1.1#53(192.168.1.1)
;; WHEN: Tue Nov 12 11:46:20 EST 2013
;; MSG SIZE rcvd: 100
As you say the cache is much faster, 2 ms as opposed to 616 ms.
But I still don't know what DNS address it actually goes to, how do I
find that?
The DNS that is in your router is what IP it goes to, which in most
cases your Internet Provider .
In /etc/resolve.conf I put the Google DNS 8.8.8.8 and 8.8.0.0 and get
better results.
You can goto http://getip.com and it will tell you what your IP is.
Hech it will even show you and everybody where you live.
11 a.m.
On 12/11/13 09:41, Tim wrote:
I get about a quarter or third of that time if I try other DNS
servers.
My own took 1500 mS the first time, then 6 ms for subsequent attempts
(local cache is much quicker than internet propagation delays). Queries
for other addresses took about 100 mS, so the BBC query must have been
answered by something further away. My server checks it cache, then (if
it doesn't have an answer) goes out to the root servers (to find out who
would know the answer), then whatever servers the roots tell it, like a
proper traditional DNS server.
I consider 1.5 secs for a response to be unusually long. I couldn't put
up with that, or worse, consistently.
I changed the router configuration to use
208.67.229.220:5353 and
208.67.222.222:5353 and see the following:
[root@box10 bobg]# dig www.bbc.co.uk
; <<>> DiG 9.9.3-rl.13207.22-P2-RedHat-9.9.3-5.P2.fc19 <<>>
www.bbc.co.uk
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 35870
;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;www.bbc.co.uk. IN A
;; ANSWER SECTION:
www.bbc.co.uk. 176 IN CNAME www.bbc.net.uk.
www.bbc.net.uk. 182 IN A 212.58.246.93
www.bbc.net.uk. 182 IN A 212.58.246.92
;; Query time: 616 msec
;; SERVER: 192.168.1.1#53(192.168.1.1)
;; WHEN: Tue Nov 12 11:46:20 EST 2013
;; MSG SIZE rcvd: 100
As you say the cache is much faster, 2 ms as opposed to 616 ms.
But I still don't know what DNS address it actually goes to, how do I
find that?
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
7:21 p.m.
On 11/11/13 19:12, Tim wrote:
That'd be HTTP caching. They don't need to subvert DNS
records for you
to see cached websites. Unless they're doing something stupid, your web
requests are still made of the original IPs, just the results are
cached.
I have the same thing, here, on my LAN. A Squid proxy server, so that
if I have guests doing the "look at this" thing amongst themselves, or a
bunch of Windows PCs doing updates, everyone after the first query sees
the cached version.
You can try it out, and see. Find a public DNS server that you can
access on a different-than-usual port. Make a rule on your gateway that
connection attempts to your router IP and DNS port get redirected to the
external DNS server on the unusual port. It's probably possible to make
an outgoing redirection rule on the PC that your testing, itself.
As far as them optimising things, with satellite internet, there's a
prolonged propagation delay. So them doing local caching means that you
get to see cached data on this side of the satellite, rather than have
to wait for it to come through it.
Years ago I used an ISP that did that sort of thing, their service was
dreadful. Everything was late, worse than dial-up. Their crap
performance was the thing that pushed me into running my own DNS
servers. Their DNS servers were even worse than their everything else
that they did. Frequently, it could take half a minute for it to return
a result. When you consider that way too many pages are a construct of
data from here, there, and everywhere, not just the sites own service,
it could take an age to load a page.
Any service that mucks you about, and fobs you off, and leaves you
trying to resolve a problem for days on end, doesn't deserve your
custom. Especially if the problem is theirs.
As I said I don't know exactly what they are doing but I do know that
I'm not smart enough to work around it. I spent quite a bit of effort
Googling "Viasat Exede" and it seems there's a dearth of information
other than sales hype. Here's an excerpt from a interview I did find,
still not very informative:
"The teleports are associated regionally with customers, so there's no
dynamic routing being done by the satellite. "If we did routing in the
satellite, we'd be getting about 10 gigabits of throughput," Dankberg
said. "With bridging, we get more than 10 times that." Just for scale,
140 gigabits per second is 20 times the bandwidth capability of the
previously existing WildBlue satellite service. "The two largest
satellite operators have about a hundred satellites," Dankberg said,
"and between them they have the same bandwidth as this one
satellite."That sort of capacity, he says, alters the economics of
satellite broadband.
There are some things bandwidth can't overcome, and one of them is the
laws of physics. Satellite is still prone to latency, which means it's
not the best choice for applications that need a short
time-of-flight—like online gaming, for example (unless you're a big fan
of lag). "We tell people up front if they're looking for a gaming
connection, this isn't it," Dankberg conceded.
There's also a small, but noticeable delay on voice-over-IP phone calls
over Exede, but not any more than a typical wireless call. I placed a
call over a VoIP phone tethered to an Exede home unit, and the
conversation was better than cellular quality."
Source:
http://arstechnica.com/business/2012/01/how-viasats-exede-makes-satellite...
He makes it sound that it's essential to stick with their software, some
parts of which are apparently contained in the modem. As I said I am
quite happy with the service, it normally works well and there is
nothing better available to me. I also recently subscribed to their voip
service and find it superior to the cell phone [we discontinued the
land-line long ago] there is usually no detectable delay due to transit
time to the satellite. My only problem was the intermittent DNS. It has
not failed today that I am aware of. The kids are not happy due to it's
inability to handle on-line gaming with their brother in Chicago. That
is not a requirement in my book.
If anyone understands what they are doing and can explain it to me I am
interested.
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
7:53 p.m.
On Sun, Nov 10, 2013 at 5:41 PM, Bob Goodwin ~ Zuni, Virginia, USA
<bobgoodwin(a)wildblue.net> wrote:
On 10/11/13 19:28, Sam Varshavchik wrote:
> Bob Goodwin ~ Zuni, Virginia, USA writes:
>> No, it won't allow any other DNS server, no matter what I configure it
>> for it always goes to theirs, that's why I gave up open.dns. This was a
>> change
>
> DD-WRT does not give you the option to use fixed DNS servers, instead of
> whatever DHCP servers it gets via DHCP from your ISP?
>
> I'm surprised to hear that. Or, rereading what you earlier wrote, maybe
> you're saying that your ISP blocks port 53 traffic, and will only let you
> use their DNS servers.
DD-WRT will let me use any DNS I want, it's Viasat's system that changes it.
They do some sort of caching to reduce traffic through the satellite
apparently ... Initially I had it configured for open.dns but it wasn't long
before I found it wasn't going there.
Wild guess: maybe the DNS information you have manually provided to
DD-WRT got made wrong by changes in the network (seems plausible given
the IP lookup data others have given), and something's going wrong
with the magic they use to override other DNS servers besides theirs
(including the one you're using which you thought was valid but
perhaps has now changed).
Try reconfiguring DD-WRT to get its DNS information from your ISP's
DHCP server and see if your problem persists. (I believe you just set
Static DNS to all 0s in DD-WRT to accomplish this.)
-T.C.
6:12 p.m.
Tim:
> Unless you're doing something odd, the amount of traffic from
DNS
> data is minuscule compared to everything else.
Bob Goodwin:
But what if they are caching stuff, e.g. foxnews, some popular video
clips, etc. and delivering them to the user without going through the
satellite loop? I don't know what they are doing but they claim to be
"optimizing" the system with their caching.
That'd be HTTP caching. They don't need to subvert DNS records for you
to see cached websites. Unless they're doing something stupid, your web
requests are still made of the original IPs, just the results are
cached.
I have the same thing, here, on my LAN. A Squid proxy server, so that
if I have guests doing the "look at this" thing amongst themselves, or a
bunch of Windows PCs doing updates, everyone after the first query sees
the cached version.
You can try it out, and see. Find a public DNS server that you can
access on a different-than-usual port. Make a rule on your gateway that
connection attempts to your router IP and DNS port get redirected to the
external DNS server on the unusual port. It's probably possible to make
an outgoing redirection rule on the PC that your testing, itself.
As far as them optimising things, with satellite internet, there's a
prolonged propagation delay. So them doing local caching means that you
get to see cached data on this side of the satellite, rather than have
to wait for it to come through it.
Years ago I used an ISP that did that sort of thing, their service was
dreadful. Everything was late, worse than dial-up. Their crap
performance was the thing that pushed me into running my own DNS
servers. Their DNS servers were even worse than their everything else
that they did. Frequently, it could take half a minute for it to return
a result. When you consider that way too many pages are a construct of
data from here, there, and everywhere, not just the sites own service,
it could take an age to load a page.
Any service that mucks you about, and fobs you off, and leaves you
trying to resolve a problem for days on end, doesn't deserve your
custom. Especially if the problem is theirs.
--
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686
Don't send private replies to my address, the mailbox is ignored. I
read messages from the public lists.
8:35 a.m.
On 11/11/13 08:40, Sam Varshavchik wrote:
Don't exclude the possibility that the modem itself is flipping out.
My DSL started going down once a week or so, some years ago.
Powercycling the modem always put me back online. Troubleshooting with
the ISP wasn't very productive. I ended up buying a new modem, and
that was it. The old modem is still sitting in a dusty cabinet, and I
actually pulled it out, in a pinch, when I suspected a problem with my
current modem.
I would not hesitate to try a new one if it was readily available to me
but as I understandit the modem is unique to their service/system. The
old one had transmit and receive coax cables between it and the antenna,
this one uses only one coax but I believe part of the radio circuitry is
in the modem with intermediate frequency signaling between it and the
antenna. It needs to do things like control the subscriber transmitter
power with varying weather conditions, stuff that would be different
than a cable modem.
Tech support did suggest that it could be a modem problem but I
generally try to avoid going through their troubleshooting procedure, at
least leave that 'til last..
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
6:18 a.m.
Allegedly, on or about 10 November 2013, Bob Goodwin ~ Zuni, Virginia,
USA sent:
when I began using this service I was advised not to try to work
around their system since it would result in my being charged for
higher usage and my allotment is limited.
Unless you're doing something odd, the amount of traffic from DNS data
is minuscule compared to everything else.
--
[tim@localhost ~]$ uname -rsvp
Linux 3.9.10-100.fc17.x86_64 #1 SMP Sun Jul 14 01:31:27 UTC 2013 x86_64
All mail to my mailbox is automatically deleted, there is no point
trying to privately email me, I will only read messages posted to the
public lists.
George Orwell's '1984' was supposed to be a warning against tyranny, not
a set of instructions for supposedly democratic governments.
6:28 p.m.
Bob Goodwin ~ Zuni, Virginia, USA writes:
No, it won't allow any other DNS server, no matter what I
configure it for
it always goes to theirs, that's why I gave up open.dns. This was a change
DD-WRT does not give you the option to use fixed DNS servers, instead of
whatever DHCP servers it gets via DHCP from your ISP?
I'm surprised to hear that. Or, rereading what you earlier wrote, maybe
you're saying that your ISP blocks port 53 traffic, and will only let you
use their DNS servers.
7:18 p.m.
On 10/11/13 19:54, Tim wrote:
There is a chance that you could still use an outside DNS service,
if
you can access one that uses a different port than your ISP is
intercepting. This would, also, depend on you either making up TCP/IP
redirection rules on your gateway so your clients queries go to that
unusual port. Or, running your own DNS server on a PC. You'd configure
your clients to use that DNS server as a normal DNS server, and you'd
configure the DNS server to do the unusual port queries.
ISPs are notorious for having awful DNS services, particularly ones that
play silly games with forcing you to use them. So running your own DNS
server, that's totally under your control, can be quite advantageous.
The mini DNS servers in some modem/routers are awful, and have no
configuration options. Some modem/routers don't have a DNS server, they
simply act as a proxy between you and the ISP supplied DNS server IPs.
Yes,when I began using this service I was advised not to try to work
around their system since it would result in my being charged for higher
usage and my allotment is limited. Further more their system does other
things that I don't want to lose, namely the "free time" during the wee
hours when I can allow the Apple computers to connect with iCloud, etc.
DD-WRT is configured to deal with thatand as I say it all works except
for this intermittent loss of DNS that has cropped up recently.
Normally the system works well, I don't want to change it, I just need
to convince them there is a DNS problem and I think it is in their
system. This thread has provided some help in how to do that.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
6:54 p.m.
Allegedly, on or about 10 November 2013, Bob Goodwin ~ Zuni, Virginia,
USA sent:
I used to subscribe to open.dns but with this satellite system it no
longer functions and we are required to use their dns which is part of
some optimization system Viasat uses.
There is a chance that you could still use an outside DNS service, if
you can access one that uses a different port than your ISP is
intercepting. This would, also, depend on you either making up TCP/IP
redirection rules on your gateway so your clients queries go to that
unusual port. Or, running your own DNS server on a PC. You'd configure
your clients to use that DNS server as a normal DNS server, and you'd
configure the DNS server to do the unusual port queries.
ISPs are notorious for having awful DNS services, particularly ones that
play silly games with forcing you to use them. So running your own DNS
server, that's totally under your control, can be quite advantageous.
The mini DNS servers in some modem/routers are awful, and have no
configuration options. Some modem/routers don't have a DNS server, they
simply act as a proxy between you and the ISP supplied DNS server IPs.
--
[tim@localhost ~]$ uname -rsvp
Linux 3.9.10-100.fc17.x86_64 #1 SMP Sun Jul 14 01:31:27 UTC 2013 x86_64
All mail to my mailbox is automatically deleted, there is no point
trying to privately email me, I will only read messages posted to the
public lists.
George Orwell's '1984' was supposed to be a warning against tyranny, not
a set of instructions for supposedly democratic governments.
6:53 p.m.
Bob Goodwin ~ Zuni, Virginia, USA writes:
On 10/11/13 19:12, Steven Stern wrote:
> Use 8.8.8.8 and 8.8.4.4 (Google Public DNS Servers) or 208.67.222.222
> and 208.67.220.220 (OpenDNS)
>
> The problem is probably with your ISP's DNS servers.
Yes, those worked well when I was able to use them, this system goes to it's
own DNS no matter what I ask it, and the DNS they provide works well most of
the time, it's just that recently I have been seeing it fail and fixing it
requires rebooting the Viasat modem.
I need to convince them they have a problem, will check more next time it
goes out.
Get someone on the horn who has a minimum of an understanding, and tell them
you can't ping their DNS servers.
6:52 p.m.
Bob Goodwin ~ Zuni, Virginia, USA writes:
On 10/11/13 19:28, Sam Varshavchik wrote:
> Bob Goodwin ~ Zuni, Virginia, USA writes:
>
>> No, it won't allow any other DNS server, no matter what I configure it for
>> it always goes to theirs, that's why I gave up open.dns. This was a change
>
> DD-WRT does not give you the option to use fixed DNS servers, instead of
> whatever DHCP servers it gets via DHCP from your ISP?
>
> I'm surprised to hear that. Or, rereading what you earlier wrote, maybe
> you're saying that your ISP blocks port 53 traffic, and will only let you
> use their DNS servers.
>
>
DD-WRT will let me use any DNS I want, it's Viasat's system that changes it.
They do some sort of caching to reduce traffic through the satellite
apparently ... Initially I had it configured for open.dns but it wasn't long
before I found it wasn't going there.
The only option I can think of is to use an SSL-based VPN tunnel, to a
friendly VPS provider. Of course, this is going to kill your bandwidth.
6:33 p.m.
On 10/11/13 19:12, Steven Stern wrote:
Use 8.8.8.8 and 8.8.4.4 (Google Public DNS Servers) or
208.67.222.222
and 208.67.220.220 (OpenDNS)
The problem is probably with your ISP's DNS servers.
Yes, those worked well when I was able to use them, this system goes to
it's own DNS no matter what I ask it, and the DNS they provide works
well most of the time, it's just that recently I have been seeing it
fail and fixing it requires rebooting the Viasat modem.
I need to convince them they have a problem, will check more next time
it goes out.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
6:47 p.m.
Allegedly, on or about 10 November 2013, Bob Goodwin ~ Zuni, Virginia,
USA sent:
Am I missing some point here? Is there some connection between dhcp
and dns that I am not aware of. My dhcp server is in the router and
deals with about thirty addresses on the LAN and works faultlessly as
far as I can see.
(Almost) only as far as making sure that your DHCP server tells its
clients the right IP to use for their DNS queries. Which it looks like
it is, from subsequent replies.
DCHP servers tell the clients various things to configure their network,
one of which is what DNS server IPs to use. Others are the clients own
IP, and things like the gateway (where the clients connect through to
the outside world), netmask (which sets the boundary between what's
inside or outside).
So, if your clients have proper networking addresses, you can ignore
debugging your DHCP server configuration. If they have some details
wrong, then you need to check whether your DHCP server is doling out the
right information, or your clients are ignoring it (some people set
manual overriding configurations on their PCs).
--
[tim@localhost ~]$ uname -rsvp
Linux 3.9.10-100.fc17.x86_64 #1 SMP Sun Jul 14 01:31:27 UTC 2013 x86_64
All mail to my mailbox is automatically deleted, there is no point
trying to privately email me, I will only read messages posted to the
public lists.
George Orwell's '1984' was supposed to be a warning against tyranny, not
a set of instructions for supposedly democratic governments.
6:41 p.m.
On 10/11/13 19:28, Sam Varshavchik wrote:
Bob Goodwin ~ Zuni, Virginia, USA writes:
> No, it won't allow any other DNS server, no matter what I configure
> it for it always goes to theirs, that's why I gave up open.dns. This
> was a change
DD-WRT does not give you the option to use fixed DNS servers, instead
of whatever DHCP servers it gets via DHCP from your ISP?
I'm surprised to hear that. Or, rereading what you earlier wrote,
maybe you're saying that your ISP blocks port 53 traffic, and will
only let you use their DNS servers.
DD-WRT will let me use any DNS I want, it's Viasat's system that changes
it. They do some sort of caching to reduce traffic through the satellite
apparently ... Initially I had it configured for open.dns but it wasn't
long before I found it wasn't going there.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
6:36 p.m.
On 10/11/13 18:24, Ed Greshko wrote:
Quite some time ago it was determined that your ISP uses
"transparent DNS proxy" to force you to use their DNS servers.
But, I really would ask them about their use of IP addresses.... as this is the output
of whois 182.63.128.68
inetnum: 182.62.0.0 - 182.63.255.255
netname: DIGI-MY
descr: DiGi Telecommunications Sdn Bhd
descr: Lot 10, Jalan Delima 1/1, Subang Hi-Tech Industrial Park,
descr: 40000 Shah Alam, Selangor Darul Ehsan, Malaysia
country: MY
admin-c: DIA1-AP
tech-c: DI39-AP
mnt-by: APNIC-HM
mnt-lower: MAINT-MY-DIGI-SB
mnt-routes: MAINT-MY-DIGI-SB
status: ALLOCATED PORTABLE
remarks: -+-+-+-+-+-+-+-+-+-+-+-++-+-+-+-+-+-+-+-+-+-+-+-+-+-+
remarks: This object can only be updated by APNIC hostmasters.
remarks: To update this object, please contact APNIC
remarks: hostmasters and include your organisation's account
remarks: name in the subject line.
remarks: -+-+-+-+-+-+-+-+-+-+-+-++-+-+-+-+-+-+-+-+-+-+-+-+-+-+
mnt-irt: IRT-DIGI-MY
changed:hm-changed@apnic.net 20100224
source: APNIC
One wonders if their screwing with IP addresses doesn't fail at times. Maybe,
sometimes, they actually transmit the DNS request to Malaysia. I would ask the ISP what
the IP address of their DNS servers should be.....
I don't remember how I got those DNS addresses but they have been
working for a couple of years without a problem until recently. I will
look into it some more tomorrow.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
6:12 p.m.
On 11/10/2013 02:29 PM, Bob Goodwin ~ Zuni, Virginia, USA wrote:
DD-WRT configuration is set for the following:
182.63.128.68 and 182.63.128.69
Use 8.8.8.8 and 8.8.4.4 (Google Public DNS Servers) or 208.67.222.222
and 208.67.220.220 (OpenDNS)
The problem is probably with your ISP's DNS servers.
--
-- Steve
5:24 p.m.
On 11/11/13 06:32, Bob Goodwin ~ Zuni, Virginia, USA wrote:
No, it won't allow any other DNS server, no matter what I
configure it for it always goes to theirs, that's why I gave up open.dns. This was a
change when they began to offer this high speed service, before that I always used some
other DNS server.
[bobg@box10 ~]$ dig google.com @182.63.128.68
; <<>> DiG 9.9.3-rl.13207.22-P2-RedHat-9.9.3-5.P2.fc19 <<>>
google.com @182.63.128.68
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 61788
;; flags: qr rd ra; QUERY: 1, ANSWER: 11, AUTHORITY: 0, ADDITIONAL: 0
;; QUESTION SECTION:
;google.com. IN A
;; ANSWER SECTION:
google.com. 1 IN A 74.125.224.206
google.com. 1 IN A 74.125.224.192
google.com. 1 IN A 74.125.224.193
google.com. 1 IN A 74.125.224.194
google.com. 1 IN A 74.125.224.195
google.com. 1 IN A 74.125.224.196
google.com. 1 IN A 74.125.224.197
google.com. 1 IN A 74.125.224.198
google.com. 1 IN A 74.125.224.199
google.com. 1 IN A 74.125.224.200
google.com. 1 IN A 74.125.224.201
;; Query time: 5 msec
;; SERVER: 182.63.128.68#53(182.63.128.68)
;; WHEN: Sun Nov 10 17:28:41 EST 2013
;; MSG SIZE rcvd: 204
I'll try dig next time it goes out. I should have done that this morning when it went
down.
Thanks for the help.
Quite some time ago it was determined that your ISP uses "transparent DNS proxy"
to force you to use their DNS servers.
But, I really would ask them about their use of IP addresses.... as this is the output of
whois 182.63.128.68
inetnum: 182.62.0.0 - 182.63.255.255
netname: DIGI-MY
descr: DiGi Telecommunications Sdn Bhd
descr: Lot 10, Jalan Delima 1/1, Subang Hi-Tech Industrial Park,
descr: 40000 Shah Alam, Selangor Darul Ehsan, Malaysia
country: MY
admin-c: DIA1-AP
tech-c: DI39-AP
mnt-by: APNIC-HM
mnt-lower: MAINT-MY-DIGI-SB
mnt-routes: MAINT-MY-DIGI-SB
status: ALLOCATED PORTABLE
remarks: -+-+-+-+-+-+-+-+-+-+-+-++-+-+-+-+-+-+-+-+-+-+-+-+-+-+
remarks: This object can only be updated by APNIC hostmasters.
remarks: To update this object, please contact APNIC
remarks: hostmasters and include your organisation's account
remarks: name in the subject line.
remarks: -+-+-+-+-+-+-+-+-+-+-+-++-+-+-+-+-+-+-+-+-+-+-+-+-+-+
mnt-irt: IRT-DIGI-MY
changed: hm-changed(a)apnic.net 20100224
source: APNIC
One wonders if their screwing with IP addresses doesn't fail at times. Maybe,
sometimes, they actually transmit the DNS request to Malaysia. I would ask the ISP what
the IP address of their DNS servers should be.....
--
Getting tired of non-Fedora discussions and self-serving posts
4:32 p.m.
On 10/11/13 16:38, staticsafe wrote:
> DD-WRT configuration is set for the following:
> >
> >182.63.128.68 and 182.63.128.69
> >
> >Normally it works well and does its job instantly as far as I can tell.
> >
dresden ~ # dig google.com @182.63.128.68
; <<>> DiG 9.9.3-P2 <<>> google.com @182.63.128.68
;; global options: +cmd
;; connection timed out; no servers could be reached
dresden ~ # dig google.com @182.63.128.69
; <<>> DiG 9.9.3-P2 <<>> google.com @182.63.128.69
;; global options: +cmd
;; connection timed out; no servers could be reached
That might be your problem, I would suggest testing yourself to see if
they are down for you as well.
As a temporary measure you can use Google Public DNS:
https://developers.google.com/speed/public-dns/
No, it won't allow any other
DNS server, no matter what I configure it
for it always goes to theirs, that's why I gave up open.dns. This was a
change when they began to offer this high speed service, before that I
always used some other DNS server.
[bobg@box10 ~]$ dig google.com @182.63.128.68
; <<>> DiG 9.9.3-rl.13207.22-P2-RedHat-9.9.3-5.P2.fc19 <<>>
google.com
@182.63.128.68
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 61788
;; flags: qr rd ra; QUERY: 1, ANSWER: 11, AUTHORITY: 0, ADDITIONAL: 0
;; QUESTION SECTION:
;google.com. IN A
;; ANSWER SECTION:
google.com. 1 IN A 74.125.224.206
google.com. 1 IN A 74.125.224.192
google.com. 1 IN A 74.125.224.193
google.com. 1 IN A 74.125.224.194
google.com. 1 IN A 74.125.224.195
google.com. 1 IN A 74.125.224.196
google.com. 1 IN A 74.125.224.197
google.com. 1 IN A 74.125.224.198
google.com. 1 IN A 74.125.224.199
google.com. 1 IN A 74.125.224.200
google.com. 1 IN A 74.125.224.201
;; Query time: 5 msec
;; SERVER: 182.63.128.68#53(182.63.128.68)
;; WHEN: Sun Nov 10 17:28:41 EST 2013
;; MSG SIZE rcvd: 204
I'll try dig next time it goes out. I should have done that this morning
when it went down.
Thanks for the help.
Bob
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
4:01 p.m.
On 11/11/13 05:38, staticsafe wrote:
On 11/10/2013 15:29, Bob Goodwin ~ Zuni, Virginia, USA wrote:
> I used to subscribe to open.dns but with this satellite system it no
> longer functions and we are required to use their dns which is part of
> some optimization system Viasat uses.
>
> DD-WRT configuration is set for the following:
>
> 182.63.128.68 and 182.63.128.69
>
> Normally it works well and does its job instantly as far as I can tell.
>
dresden ~ # dig google.com @182.63.128.68
; <<>> DiG 9.9.3-P2 <<>> google.com @182.63.128.68
;; global options: +cmd
;; connection timed out; no servers could be reached
dresden ~ # dig google.com @182.63.128.69
; <<>> DiG 9.9.3-P2 <<>> google.com @182.63.128.69
;; global options: +cmd
;; connection timed out; no servers could be reached
That might be your problem, I would suggest testing yourself to see if
they are down for you as well.
As a temporary measure you can use Google Public DNS:
https://developers.google.com/speed/public-dns/
whois shows those IP's belonging to DiGi Telecommunications Sdn Bhd in Malaysia.
--
Getting tired of non-Fedora discussions and self-serving posts
2:29 p.m.
On 10/11/13 15:19, staticsafe wrote:
On 11/10/2013 15:17, Bob Goodwin ~ Zuni, Virginia, USA wrote:
> Well this is what I see:
>
> [bobg@box10 ~]$ cat /etc/resolv.conf
> # Generated by NetworkManager
> nameserver 192.168.1.1
>
>
> 192.168.1.1 is my Linksys E3000 DD-WRT router.
>
DD-WRT usually runs a forwarder, check your router to see where it is
forwarding DNS queries.
I used to subscribe to open.dns but with this satellite system it no
longer functions and we are required to use their dns which is part of
some optimization system Viasat uses.
DD-WRT configuration is set for the following:
182.63.128.68 and 182.63.128.69
Normally it works well and does its job instantly as far as I can tell.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
3:38 p.m.
On 11/10/2013 15:29, Bob Goodwin ~ Zuni, Virginia, USA wrote:
I used to subscribe to open.dns but with this satellite system it no
longer functions and we are required to use their dns which is part of
some optimization system Viasat uses.
DD-WRT configuration is set for the following:
182.63.128.68 and 182.63.128.69
Normally it works well and does its job instantly as far as I can tell.
dresden ~ # dig google.com @182.63.128.68
; <<>> DiG 9.9.3-P2 <<>> google.com @182.63.128.68
;; global options: +cmd
;; connection timed out; no servers could be reached
dresden ~ # dig google.com @182.63.128.69
; <<>> DiG 9.9.3-P2 <<>> google.com @182.63.128.69
;; global options: +cmd
;; connection timed out; no servers could be reached
That might be your problem, I would suggest testing yourself to see if
they are down for you as well.
As a temporary measure you can use Google Public DNS:
https://developers.google.com/speed/public-dns/
--
staticsafe
O< ascii ribbon campaign - stop html mail - www.asciiribbon.org
Please don't top post. It is not logical.
Please don't CC me! I'm subscribed to whatever list I just posted on.
2:19 p.m.
On 11/10/2013 15:17, Bob Goodwin ~ Zuni, Virginia, USA wrote:
Well this is what I see:
[bobg@box10 ~]$ cat /etc/resolv.conf
# Generated by NetworkManager
nameserver 192.168.1.1
192.168.1.1 is my Linksys E3000 DD-WRT router.
DD-WRT usually runs a forwarder, check your router to see where it is
forwarding DNS queries.
--
staticsafe
O< ascii ribbon campaign - stop html mail - www.asciiribbon.org
Please don't top post. It is not logical.
Please don't CC me! I'm subscribed to whatever list I just posted on.
2:10 p.m.
On 11/10/2013 14:57, Bob Goodwin ~ Zuni, Virginia, USA wrote:
Recently we have begun to have problems with our browsers stopping to
work, ask it to go to a bookmark address and nothing happens, everything
stops until an error message comes up that the address can not be found.
Thunderbird continues to receive e-mail unless I restart Thunderbird,
then it is unable to find gmail.com. The problem exists on two F-19
computers and several Apple computers on the LAN, in fact I often become
aware of the problem when someone complains that the internet connection
is down.
I can always restore normal operation by re-starting the Viasat modem, a
process that takes several minutes and is an annoyance.
I have determined that I can enter a numeric address e.g. 199.106.52.212
for viasat.com and that works as expected. At the request of viasat tech
support I tried taking the router out of the circuit and connecting
directly to the modem, that effected no change. Curiously, at the time
this this morning when the problem was manifest tech support's "computer
was down" which left me wondering if there was a correlation with my
apparent DNS problem. I was asked to call back in a few hours, I did that.
The tech support person seemed to take offense at my claiming there was
a DNS problem and went on to explain that I probably have a dhcp problem
and we need to start by "re-setting my NIC." From there on the
conversation began to deteriorate and she accused me of shouting at her,
I was able to extricate myself gracefully after that and agree that I
would call back when I had more time to devote to troubleshooting.
Am I missing some point here? Is there some connection between dhcp and
dns that I am not aware of. My dhcp server is in the router and deals
with about thirty addresses on the LAN and works faultlessly as far as I
can see.
Any thoughts appreciated,
Bob
DHCP is used to distribute DNS resolvers to your LAN.
What resolvers do you get via DHCP?
(Check Network Manager or /etc/resolv.conf)
--
staticsafe
O< ascii ribbon campaign - stop html mail - www.asciiribbon.org
Please don't top post. It is not logical.
Please don't CC me! I'm subscribed to whatever list I just posted on.
2:17 p.m.
On 10/11/13 15:10, staticsafe wrote:
On 11/10/2013 14:57, Bob Goodwin ~ Zuni, Virginia, USA wrote:
> Recently we have begun to have problems with our browsers stopping to
> work, ask it to go to a bookmark address and nothing happens, everything
> stops until an error message comes up that the address can not be found.
>
> Thunderbird continues to receive e-mail unless I restart Thunderbird,
> then it is unable to find gmail.com. The problem exists on two F-19
> computers and several Apple computers on the LAN, in fact I often become
> aware of the problem when someone complains that the internet connection
> is down.
>
> I can always restore normal operation by re-starting the Viasat modem, a
> process that takes several minutes and is an annoyance.
>
> I have determined that I can enter a numeric address e.g. 199.106.52.212
> for viasat.com and that works as expected. At the request of viasat tech
> support I tried taking the router out of the circuit and connecting
> directly to the modem, that effected no change. Curiously, at the time
> this this morning when the problem was manifest tech support's "computer
> was down" which left me wondering if there was a correlation with my
> apparent DNS problem. I was asked to call back in a few hours, I did that.
>
> The tech support person seemed to take offense at my claiming there was
> a DNS problem and went on to explain that I probably have a dhcp problem
> and we need to start by "re-setting my NIC." From there on the
> conversation began to deteriorate and she accused me of shouting at her,
> I was able to extricate myself gracefully after that and agree that I
> would call back when I had more time to devote to troubleshooting.
>
> Am I missing some point here? Is there some connection between dhcp and
> dns that I am not aware of. My dhcp server is in the router and deals
> with about thirty addresses on the LAN and works faultlessly as far as I
> can see.
>
> Any thoughts appreciated,
>
> Bob
>
>
DHCP is used to distribute DNS resolvers to your LAN.
What resolvers do you get via DHCP?
(Check Network Manager or /etc/resolv.conf)
Well this is what I see:
[bobg@box10 ~]$ cat /etc/resolv.conf
# Generated by NetworkManager
nameserver 192.168.1.1
192.168.1.1 is my Linksys E3000 DD-WRT router.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
7:40 a.m.
Bob Goodwin ~ Zuni, Virginia, USA writes:
But what if they are caching stuff, e.g. foxnews, some popular video clips,
etc. and delivering them to the user without going through the satellite
loop? I don't know what they are doing but they claim to be "optimizing"
the
system with their caching. I'm not concerned about the trivial usage
attributable to DNS. I just want it to work. As it has been I've had to
restart the Viasat modem several times a week, an annoyance I will have to
put up with it seems, my experience has been that such things eventually
change, hopefully for the better.
Don't exclude the possibility that the modem itself is flipping out. My DSL
started going down once a week or so, some years ago. Powercycling the modem
always put me back online. Troubleshooting with the ISP wasn't very
productive. I ended up buying a new modem, and that was it. The old modem is
still sitting in a dusty cabinet, and I actually pulled it out, in a pinch,
when I suspected a problem with my current modem.
12:36 p.m.
On 12/11/13 13:28, Jim wrote:
On 11/12/2013 01:19 PM, Bob Goodwin ~ Zuni, Virginia, USA wrote:
>
>
> On 12/11/13 13:03, Jim wrote:
>> The DNS that is in your router is what IP it goes to, which in most
>> cases your Internet Provider .
>
> Yes, normally you would expect that but Viasat Exede forces you to
> their DNS no matter what you set your router for and there is nothing
> wrong with the way that works as long as it is working normally. I
> have had some intermittent problems recently but since my complaint
> yesterday it has worked well. Coincidence perhaps?
>
> Bob
>
The DNS numbers you have in /etc/resolv.conf would be the DNS-IP that
it would goto.
At least that is my understanding.
It goes to the router, 192.168.1.1 and presently the router is set to
use open.dns.
[bobg@box10 ~]$ cat /etc/resolv.conf
# Generated by NetworkManager
nameserver 192.168.1.1
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
6:17 p.m.
On 11Nov2013 20:21, Bob Goodwin <bobgoodwin(a)wildblue.net> wrote:
[...]
Source:
http://arstechnica.com/business/2012/01/how-viasats-exede-makes-satellite...
He makes it sound that it's essential to stick with their software,
some parts of which are apparently contained in the modem. As I said
I am quite happy with the service, it normally works well and there
is nothing better available to me. [...]
If anyone understands what they are doing and can explain it to me I
am interested.
If it is like the setup in our Gilat satellite modem here, there are two
primary things going on. (Based on reading scattered stuff on the net
when trying to debug an annoying problem some months ago.)
The modem/satellite system does HTTP prefetching; when you fetch a
web page the traffic is sniffed and requests for the page content
(images etc) are prefetched. This step may happen upstream of the
modem (eg in a proxy at the ISP) or it may happen in the modem -
this is unclear; that "HTTP acceleration" is a modem-side setting
suggests it may be happening in the modem.
The other thing the Gilat modems do is "TCP acceleration". I gather
a TCP connection is terminated localy at the modem, and the data
stream sent with a more efficient protocol over the satellite section
of the link and a TCP connection established on the upstream (far
side of you->satellite->downstation) to carry the traffic from there
to the target server. The TCP handshake takes place in real time -
the modem does not falsely complete a connection only to have
upstream fail, but past that the traffic is proxied.
This may sound ineffectual, but in fact it has throughput benefits.
The TCP connection from upstream to the target server can transfer
data rapidly across the fast and low latency ground based network.
The data beamed to/from the satellite is sent directly using the
satellite specific data integrity protocols without the TCP packet
tracking protocol, and the data from your host to the modem travels
with low round trip latency to/from the modem
I'm speaking here of the TCP packet ACK stuff, which is now you<->modem
and upstream<->target-server.
This results in greater throughput.
Cheers,
--
Cameron Simpson <cs(a)zip.com.au>
I took that Reading Dynamics course, and it really works. I read _War and
Peace_ in an hour last night. It's about Russia. - W. Allen, ca. 1962
12:26 a.m.
On Wed, 2013-11-13 at 11:03 +1100, Cameron Simpson wrote:
Maybe, maybe not. If one of their DNS servers was flakey, your
complaint would not be alone. If they replaced/repaired their server,
everyone will see the fix.
And the front line phone staff may be unaware of the issue.
It depends on the ISP. I can mention two major ones in this country,
Optus and Telstra, who don't just service customers, but are backbones
for most other smaller ISPs, who will flatly refuse to admit that
there's ever anything wrong with their equipment, will endlessly tell
you reset your equipment, despite all evidence to the contrary that
their system is up the creek - e.g. their support newsgroup is full of
customers complaining that the DNS server isn't working. So I am as
dismissive of anything they say, as they are of their customer
complaints (they're as arrogant as that against their phone customers,
as well).
I could mention an older ISP that doesn't exist anymore, Picknowl, that
would get back to your support query and say they've checked their
equipment, noticed something wrong and reset it, or didn't find a
problem but reset it just in case. And, back in the dialup days, if you
had trouble connecting, they'd ask you to dial a specific number, and
they'd watch their terminal to check what your system was trying to do
as it logged in. I had more faith in them. Unfortunately, they got
bought out and ruined.
Another ISP (Adam) seems to publish their internal fault reports on
their status pages. You can see fault history, scheduled maintenance,
reported faults, and expected repair times. It's usually kept up to
date quite promptly.
But if you're going to make a fault report to an ISP, with any hope of
getting a proper response, you need to make a proper report that they
can follow up on. Mention your own IP, addresses you've tried to access
but failed, your time and timezone - so they can check their logs, not
just their equipment at the current moment.
--
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686
Don't send private replies to my address, the mailbox is ignored. I
read messages from the public lists.
11:59 p.m.
On Tue, 2013-11-12 at 23:09 +0800, Ed Greshko wrote:
If the ISP is doing transparent DNS proxy I seriously doubt that any
of that matters.
It depends on how they're doing it. If they simply intercept port 53
traffic, then not using port 53 for your queries will bypass it. But if
they also snoop on some other ports, or snoop on all traffic for things
that look like DNS queries, then you're screwed.
My opinion of ISPs that do transparent proxying, is rather low. I
understand the point of view that some do that, so users don't have to
configure their software. But my experience has been that ISPs do it
to: Lower their operating costs, while still charging you the same, and
to rake in more customers than their equipment (otherise) has bandwidth
for.
--
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686
Don't send private replies to my address, the mailbox is ignored. I
read messages from the public lists.
5:58 p.m.
On 10Nov2013 20:18, Bob Goodwin <bobgoodwin(a)wildblue.net> wrote:
[...] Further more their system
does other things that I don't want to lose, namely the "free time"
during the wee hours when I can allow the Apple computers to connect
with iCloud, [...]
Regarding reigning in the Apple stuff, what approach are you using here?
Cheers,
--
Cameron Simpson <cs(a)zip.com.au>
A program in conformance will not tend to stay in conformance, because even if
it doesn't change, the standard will. - Norman Diamond <diamond(a)jit.dec.com>
12:30 p.m.
On 11/12/2013 10:03 AM, Jim wrote:
You can goto http://getip.com and it will tell you what your IP is. Hech
it will even show you and everybody where you live.
Actually, no. It won't. What it shows is where the DSLAM you're
connected to is, and in my case, that's several miles away.
9:09 a.m.
On 11/12/13 22:41, Tim wrote:
Next time it happens, start playing with the dig tool. Make queries
against your router, and against the DNS IPs the ISP provides (to the
router), and try some external DNS servers that have different ports
(the dig command lets you specify ports, read the man page, brief
outline below). Do queries for different addresses, if you keep
querying the same address, you'll just get results from a cache.
If the ISP is doing transparent DNS proxy I seriously doubt that any of that matters.
--
Getting tired of non-Fedora discussions and self-serving posts
7:24 a.m.
On 11/11/13 07:18, Tim wrote:
Allegedly, on or about 10 November 2013, Bob Goodwin ~ Zuni,
Virginia,
USA sent:
> when I began using this service I was advised not to try to work
> around their system since it would result in my being charged for
> higher usage and my allotment is limited.
Unless you're doing something odd, the amount of traffic from DNS data
is minuscule compared to everything else.
But what if they are caching stuff, e.g. foxnews, some popular video
clips, etc. and delivering them to the user without going through the
satellite loop? I don't know what they are doing but they claim to be
"optimizing" the system with their caching. I'm not concerned about the
trivial usage attributable to DNS. I just want it to work. As it has
been I've had to restart the Viasat modem several times a week, an
annoyance I will have to put up with it seems, my experience has been
that such things eventually change, hopefully for the better.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
7:05 p.m.
On 10/11/13 19:53, Sam Varshavchik wrote:
>
> I need to convince them they have a problem, will check more next
> time it goes out.
Get someone on the horn who has a minimum of an understanding, and
tell them you can't ping their DNS servers.
Yes, I intend to do that, thought I had enough information this morning
but the tech support person took offense at being told their DNS was
failing and began telling me some bazaar things, she was confusing mac
addresses with URLs, wanted to test my NIC. I explained that the trouble
was with two Linux computers and several Apple Mac devices. I was
getting no where ... :-(
I will try again when the problem reoccurs.
--
http://www.qrz.com/db/W2BOD
box10 Fedora-19/64 bit Linux/XFCE
8:41 a.m.
Allegedly, on or about 11 November 2013, Bob Goodwin sent:
He makes it sound that it's essential to stick with their
software,
some parts of which are apparently contained in the modem. As I said I
am quite happy with the service, it normally works well and there is
nothing better available to me. I also recently subscribed to their
voip service and find it superior to the cell phone [we discontinued
the land-line long ago] there is usually no detectable delay due to
transit time to the satellite. My only problem was the intermittent
DNS.
Looking at that page, from what I can gather. They have a ground based
network, with high speed and high bandwidth. It connects to you, other
users, and probably other local services. The satellite is their link
to things further away, and has latency by still high bandwidth, so
things work well enough by the time that they've started and filled up a
cache (e.g. watching streaming video). The modem has some custom
compression to speed thing up even more - though any old BBS SysOp can
tell you that you can only speed up still-compressible data, that way,
not all data.
So, again, I say just find yourself a decent DNS server, and use it
instead. You'll still be going through their caches and proxies for the
data, you'll just be getting the addresses quicker.
Next time it happens, start playing with the dig tool. Make queries
against your router, and against the DNS IPs the ISP provides (to the
router), and try some external DNS servers that have different ports
(the dig command lets you specify ports, read the man page, brief
outline below). Do queries for different addresses, if you keep
querying the same address, you'll just get results from a cache.
dig @server -p number address
Substituting "server" for the address of server you're going to query,
"number" for the port number you're querying through (if not the usual
port 53), and address for the address you want to query.
Have you tried OpenDNS on port 5353 instead of port 53? If you're
lucky, your ISP is only doing a simplistic buggery of DNS on 53.
e.g. dig @208.67.222.222 -p 5353 www.bbc.co.uk
; <<>> DiG 9.9.3-rl.156.01-P1-RedHat-9.9.3-3.P1.fc17 <<>>
@208.67.222.222 -p 5353 www.bbc.co.uk
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 61634
;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;www.bbc.co.uk. IN A
;; ANSWER SECTION:
www.bbc.co.uk. 291 IN CNAME www.bbc.net.uk.
www.bbc.net.uk. 165 IN A 212.58.244.68
www.bbc.net.uk. 165 IN A 212.58.244.69
;; Query time: 98 msec
;; SERVER: 208.67.222.222#5353(208.67.222.222)
;; WHEN: Wed Nov 13 01:03:31 CST 2013
;; MSG SIZE rcvd: 100
I get about a quarter or third of that time if I try other DNS servers.
My own took 1500 mS the first time, then 6 ms for subsequent attempts
(local cache is much quicker than internet propagation delays). Queries
for other addresses took about 100 mS, so the BBC query must have been
answered by something further away. My server checks it cache, then (if
it doesn't have an answer) goes out to the root servers (to find out who
would know the answer), then whatever servers the roots tell it, like a
proper traditional DNS server.
I consider 1.5 secs for a response to be unusually long. I couldn't put
up with that, or worse, consistently.
--
[tim@localhost ~]$ uname -rsvp
Linux 3.9.10-100.fc17.x86_64 #1 SMP Sun Jul 14 01:31:27 UTC 2013 x86_64
All mail to my mailbox is automatically deleted, there is no point
trying to privately email me, I will only read messages posted to the
public lists.
George Orwell's '1984' was supposed to be a warning against tyranny, not
a set of instructions for supposedly democratic governments.
3809
days inactive
3813
days old
43 comments
11 participants
participants (11)
-
Bob Goodwin -
Cameron Simpson -
Ed Greshko -
Joe Zeff -
Mickey -
Rick Stevens -
Sam Varshavchik -
staticsafe -
SternData -
T.C. Hollingsworth -
Tim