PPPoA failures question

I support several sites which have a common failure. Details are:

ISP = Demon
Router = Various versions of Vigor (2600plus/i : 2600plus : 2600V : 2800)
VPN = LAN-to-LAN, configured for "always on", connecting to another Vigor
router

About once every 6 weeks the router loses its internet connection.
Inspection of the router (necessitating a site visit) shows that the ADSL
sync is good, and reasonable figures are shown for up/down speed, SNR margin
and attenuation. But the WAN status shows the PPPoA link is down. Clicking
the button to "Dial PPPoA" has no effect.

Rebooting the router, or unplugging and reconnecting the phone line, will
always restore the conection. However the problem (by Sod's law) seems only
to occur then the sites are unattended (as in over a weekend).

On the one occasion that I struggled to get through to Demon's technical
support they appeared to suggest that the router was lying, in that their
system showed no ADSL sync; but I've little confidence in that.

I have similar sites where the only difference is the ISP but these sites
probably make less use of the VPN. I don't have any Demon sites where there
is no use of any VPN, so can't speculate as to whether the VPN is a
contributory factor.

Has anybody seen anything similar? If so, have you a resolution?

Thanks,


--
Graham J

Graham J wrote:
> I support several sites which have a common failure. Details are:
>
> ISP = Demon
> Router = Various versions of Vigor (2600plus/i : 2600plus : 2600V : 2800)
> VPN = LAN-to-LAN, configured for "always on", connecting to another Vigor
> router
>
> About once every 6 weeks the router loses its internet connection.
> Inspection of the router (necessitating a site visit) shows that the ADSL
> sync is good, and reasonable figures are shown for up/down speed, SNR margin
> and attenuation. But the WAN status shows the PPPoA link is down. Clicking
> the button to "Dial PPPoA" has no effect.
>
> Rebooting the router, or unplugging and reconnecting the phone line, will
> always restore the conection. However the problem (by Sod's law) seems only
> to occur then the sites are unattended (as in over a weekend).
>
> On the one occasion that I struggled to get through to Demon's technical
> support they appeared to suggest that the router was lying, in that their
> system showed no ADSL sync; but I've little confidence in that.
>
> I have similar sites where the only difference is the ISP but these sites
> probably make less use of the VPN. I don't have any Demon sites where there
> is no use of any VPN, so can't speculate as to whether the VPN is a
> contributory factor.
>
> Has anybody seen anything similar? If so, have you a resolution?
>
> Thanks,
>
>
I've seen similar, bit not the same..


Once it was endemic..ATM failures..BT capacity problem caused by major
interconnect failure.

Never admitted by anyone but inferred from a statment by NBT sometime later.

Another time, similar, except could not get any ATM to ISP. Never
admitted but cleared when ISP 'upgraded firmware in DSL termination
equipment'

Graham J wrote:

> Has anybody seen anything similar? If so, have you a resolution?

Have a look at the thread "So who knows what "Circuit has been re-mapped on
the telnet" means?". Might be the same thing. Might not.


--
<http://ale.cx/> (AIM:troffasky) ((E-Mail Removed))
21:53:00 up 83 days, 4 min, 3 users, load average: 0.00, 0.03, 0.03
Sexy ladies, and nasty boys, all freaky freakin', to the robot noise

"alexd" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> Graham J wrote:
>
>> Has anybody seen anything similar? If so, have you a resolution?
>
> Have a look at the thread "So who knows what "Circuit has been re-mapped
> on
> the telnet" means?". Might be the same thing. Might not.

The reply from Gordon Henderson suggests that his situation is permanent.
His implication is that changing the router (which involves breaking the
phone connection) is not sufficient to restore service. His analysis
suggests the exchange does not send LLC packets - ever!

By contrast, my failures can be cured by breaking / reconnecting the 'phone
line, which suggests that the exchange equipment has simply got stuck, and
is reset when sync is lost and then restored.

So I don't think its the same. Also, the failures I see are at times of
likely high use (6pm on a Friday) rather than the early hours when
maintenance work might reconfigure BTs network.

--
Graham J

Graham J wrote:
> "alexd" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
>> Graham J wrote:
>>
>>> Has anybody seen anything similar? If so, have you a resolution?
>> Have a look at the thread "So who knows what "Circuit has been re-mapped
>> on
>> the telnet" means?". Might be the same thing. Might not.
>
> The reply from Gordon Henderson suggests that his situation is permanent.
> His implication is that changing the router (which involves breaking the
> phone connection) is not sufficient to restore service. His analysis
> suggests the exchange does not send LLC packets - ever!
>
> By contrast, my failures can be cured by breaking / reconnecting the 'phone
> line, which suggests that the exchange equipment has simply got stuck, and
> is reset when sync is lost and then restored.
>
> So I don't think its the same. Also, the failures I see are at times of
> likely high use (6pm on a Friday) rather than the early hours when
> maintenance work might reconfigure BTs network.
>
Yup.

I am sensing that BT's backhaul is actual near capacity, or that some
operators are pushing it that way.

Lost packets there could easily account for this stuff.
Double check you have the right configs in you router as sometimes they
work, but flakily, with other than optimal settings.




You should mave PPOA VC MUX with VPI 0 and VCI 38 I think. And be using
CHAP ..

"The Natural Philosopher" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> Graham J wrote:
>> "alexd" <(E-Mail Removed)> wrote in message
>> news:(E-Mail Removed)...
>>> Graham J wrote:
>>>
>>>> Has anybody seen anything similar? If so, have you a resolution?
>>> Have a look at the thread "So who knows what "Circuit has been re-mapped
>>> on
>>> the telnet" means?". Might be the same thing. Might not.
>>
>> The reply from Gordon Henderson suggests that his situation is permanent.
>> His implication is that changing the router (which involves breaking the
>> phone connection) is not sufficient to restore service. His analysis
>> suggests the exchange does not send LLC packets - ever!
>>
>> By contrast, my failures can be cured by breaking / reconnecting the
>> 'phone line, which suggests that the exchange equipment has simply got
>> stuck, and is reset when sync is lost and then restored.
>>
>> So I don't think its the same. Also, the failures I see are at times of
>> likely high use (6pm on a Friday) rather than the early hours when
>> maintenance work might reconfigure BTs network.
>>
> Yup.
>
> I am sensing that BT's backhaul is actual near capacity, or that some
> operators are pushing it that way.
>
> Lost packets there could easily account for this stuff.
> Double check you have the right configs in you router as sometimes they
> work, but flakily, with other than optimal settings.

You mean
VPI = 0
VCI = 38
Type = VC MUX
Protocol = PPPoA
Modulation = G.DMT
PPP authentication = PAP or CHAP

???

The Vigor router can be configured to pick most of these up from the ISP.
If LLC/SNAP or PPPoE were required would you expect good reliability with
the wrong settings, and only the once in 6 weeks failures I've seen?

Graham J wrote:
> "The Natural Philosopher" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
>> Graham J wrote:
>>> "alexd" <(E-Mail Removed)> wrote in message
>>> news:(E-Mail Removed)...
>>>> Graham J wrote:
>>>>
>>>>> Has anybody seen anything similar? If so, have you a resolution?
>>>> Have a look at the thread "So who knows what "Circuit has been re-mapped
>>>> on
>>>> the telnet" means?". Might be the same thing. Might not.
>>> The reply from Gordon Henderson suggests that his situation is permanent.
>>> His implication is that changing the router (which involves breaking the
>>> phone connection) is not sufficient to restore service. His analysis
>>> suggests the exchange does not send LLC packets - ever!
>>>
>>> By contrast, my failures can be cured by breaking / reconnecting the
>>> 'phone line, which suggests that the exchange equipment has simply got
>>> stuck, and is reset when sync is lost and then restored.
>>>
>>> So I don't think its the same. Also, the failures I see are at times of
>>> likely high use (6pm on a Friday) rather than the early hours when
>>> maintenance work might reconfigure BTs network.
>>>
>> Yup.
>>
>> I am sensing that BT's backhaul is actual near capacity, or that some
>> operators are pushing it that way.
>>
>> Lost packets there could easily account for this stuff.
>> Double check you have the right configs in you router as sometimes they
>> work, but flakily, with other than optimal settings.
>
> You mean
> VPI = 0
> VCI = 38
> Type = VC MUX
> Protocol = PPPoA
> Modulation = G.DMT
> PPP authentication = PAP or CHAP
>
> ???

Probably ;-)
On my router type and protcocol are munged together

>
> The Vigor router can be configured to pick most of these up from the ISP.
> If LLC/SNAP or PPPoE were required would you expect good reliability with
> the wrong settings, and only the once in 6 weeks failures I've seen?
>

I dunno. I have had my router misconfigured and it still worked..

I mean it SOUNDS like an ATM session failure..or a PPP session failure,
so those are the areas to look at. My knowldge is extremely rusty here,
but PAP I thought was a one time authentication, whereas PAP is a
continuous re-establishing of authentication..many ISPs accept both. One
may work better. For example.

I also eeme to remember that PPOE is allowable..but ppoA is
recommended..but again, check with your ISP.

Do you have deep level ATM diags on the router?








>
>
>

On 26 Feb, 18:28, The Natural Philosopher <a...@b.c> wrote:
> Graham J wrote:
> > "The Natural Philosopher" <a...@b.c> wrote in message
> >news:(E-Mail Removed).. .
> >> Graham J wrote:
> >>> "alexd" <troffa...@hotmail.com> wrote in message
> >>>news:(E-Mail Removed)...
> >>>> Graham J wrote:
>
> >>>>> Has anybody seen anything similar? *If so, have you a resolution?
> >>>> Have a look at the thread "So who knows what "Circuit has been re-mapped
> >>>> on
> >>>> the telnet" means?". Might be the same thing. Might not.
> >>> The reply from Gordon Henderson suggests that his situation is permanent.
> >>> His implication is that changing the router (which involves breaking the
> >>> phone connection) is not sufficient to restore service. *His analysis
> >>> suggests the exchange does not send LLC packets - ever!
>
> >>> By contrast, my failures can be cured by breaking / reconnecting the
> >>> 'phone line, which suggests that the exchange equipment has simply got
> >>> stuck, and is reset when sync is lost and then restored.
>
> >>> So I don't think its the same. *Also, the failures I see are at times of
> >>> likely high use (6pm on a Friday) rather than the early hours when
> >>> maintenance work might reconfigure BTs network.
>
> >> Yup.
>
> >> I am sensing that BT's backhaul is actual near capacity, or that some
> >> operators are pushing it that way.
>
> >> Lost packets there could easily account for this stuff.
> >> Double check you have the right configs in you router as sometimes they
> >> work, but flakily, with other than optimal settings.
>
> > You mean
> > VPI * *= 0
> > VCI * *= 38
> > Type * *= VC MUX
> > Protocol * *= PPPoA
> > Modulation = G.DMT
> > PPP authentication * *= PAP or CHAP
>
> > ???
>
> Probably ;-)
> On my router type and protcocol are munged together
>
>
>
> > The Vigor router can be configured to pick most of these up from the ISP.
> > If *LLC/SNAP or PPPoE were required would you expect good reliabilitywith
> > the wrong settings, and only the once in 6 weeks failures I've seen?
>
> I dunno. I have had my router misconfigured and it still worked..
>
> I mean it SOUNDS like an ATM session failure..or a PPP session failure,
> so those are the areas to look at. My knowldge is extremely rusty here,
> but PAP I thought was a one time authentication, whereas PAP is a
> continuous re-establishing of authentication..many ISPs accept both. One
> may work better. For example.
>
> I also eeme to remember that PPOE is allowable..but ppoA is
> recommended..but again, check with your ISP.
>
> Do you have deep level ATM diags on the router?

Graham,

Have you checked the software version in the router?
They come out with new 'better' ones all the time. In
most cases you can download from the manufacturers web site.
Check the information VERY carefully since there is
different software for different countries and many
different hardwre versions and things like that.

If you NEED this to be working every day I would suggest
considering having a spare router standing by before upgrade
attempt.

Also, I have once, just for fun tried PPPoE on BT
and it just worked. It's what I would do if I was an ISP.
i.e. support as much variety as possible in the hope of
reducing support call numbers.

I also tried:-
ansi-dmt ANSI T1.413
itu-dmt ITU G.992.1 Annex A

and they both worked too.

That's an idea - I have once had trouble with dmt-auto
and chose the one that auto did not choose and
solved a problem. Never know your luck.
Router software upgrade took that one away later on.
I assume it was a firmware bug in my router that
was causing the problem with one of the dmt thingies.
Not that I have a clue what dmt is

On auto-dmt I seem to always get the ITU one.

"bod43" <(E-Mail Removed)> wrote in message
news:5449b226-bca1-4b82-bd82-(E-Mail Removed)...
On 26 Feb, 18:28, The Natural Philosopher <a...@b.c> wrote:
> Graham J wrote:
> > "The Natural Philosopher" <a...@b.c> wrote in message
> >news:(E-Mail Removed).. .
> >> Graham J wrote:
> >>> "alexd" <troffa...@hotmail.com> wrote in message
> >>>news:(E-Mail Removed)...
> >>>> Graham J wrote:
>
> >>>>> Has anybody seen anything similar? If so, have you a resolution?
> >>>> Have a look at the thread "So who knows what "Circuit has been
> >>>> re-mapped
> >>>> on
> >>>> the telnet" means?". Might be the same thing. Might not.
> >>> The reply from Gordon Henderson suggests that his situation is
> >>> permanent.
> >>> His implication is that changing the router (which involves breaking
> >>> the
> >>> phone connection) is not sufficient to restore service. His analysis
> >>> suggests the exchange does not send LLC packets - ever!
>
> >>> By contrast, my failures can be cured by breaking / reconnecting the
> >>> 'phone line, which suggests that the exchange equipment has simply got
> >>> stuck, and is reset when sync is lost and then restored.
>
> >>> So I don't think its the same. Also, the failures I see are at times
> >>> of
> >>> likely high use (6pm on a Friday) rather than the early hours when
> >>> maintenance work might reconfigure BTs network.
>
> >> Yup.
>
> >> I am sensing that BT's backhaul is actual near capacity, or that some
> >> operators are pushing it that way.
>
> >> Lost packets there could easily account for this stuff.
> >> Double check you have the right configs in you router as sometimes they
> >> work, but flakily, with other than optimal settings.
>
> > You mean
> > VPI = 0
> > VCI = 38
> > Type = VC MUX
> > Protocol = PPPoA
> > Modulation = G.DMT
> > PPP authentication = PAP or CHAP
>
> > ???
>
> Probably ;-)
> On my router type and protcocol are munged together
>
>
>
> > The Vigor router can be configured to pick most of these up from the
> > ISP.
> > If LLC/SNAP or PPPoE were required would you expect good reliability
> > with
> > the wrong settings, and only the once in 6 weeks failures I've seen?
>
> I dunno. I have had my router misconfigured and it still worked..
>
> I mean it SOUNDS like an ATM session failure..or a PPP session failure,
> so those are the areas to look at. My knowldge is extremely rusty here,
> but PAP I thought was a one time authentication, whereas PAP is a
> continuous re-establishing of authentication..many ISPs accept both. One
> may work better. For example.
>
> I also eeme to remember that PPOE is allowable..but ppoA is
> recommended..but again, check with your ISP.
>
> Do you have deep level ATM diags on the router?

Graham,

Have you checked the software version in the router?
They come out with new 'better' ones all the time. In
most cases you can download from the manufacturers web site.
Check the information VERY carefully since there is
different software for different countries and many
different hardwre versions and things like that.

If you NEED this to be working every day I would suggest
considering having a spare router standing by before upgrade
attempt.

Also, I have once, just for fun tried PPPoE on BT
and it just worked. It's what I would do if I was an ISP.
i.e. support as much variety as possible in the hope of
reducing support call numbers.

I also tried:-
ansi-dmt ANSI T1.413
itu-dmt ITU G.992.1 Annex A

and they both worked too.

That's an idea - I have once had trouble with dmt-auto
and chose the one that auto did not choose and
solved a problem. Never know your luck.
Router software upgrade took that one away later on.
I assume it was a firmware bug in my router that
was causing the problem with one of the dmt thingies.
Not that I have a clue what dmt is

On auto-dmt I seem to always get the ITU one.

-----------------------------------------------

Reply here:

I agree I have seen PPPoE work when the connection is ADSL and one would
expect PPPoA.

At one site I replaced the 2600plus with a brand new 2800 with the most
up-to-date firmware. All I can say is that the failures occur slightly less
frequently. The 2600plus is working perfectly adequately at another site
where the ISP is not Demon - but the VPN is only used for maintenance, it's
not up continuously.

I'm beginning to suspect that is is a Vigor problem with the "always on"
LAN-to-LAN VPN that carries a lot of traffic; and by chance the only sites
where it manifests itself are those with a connection from Demon.

--
Graham J

On Wed, 25 Feb 2009 10:49:27 +0000, Graham J wrote:

> I support several sites which have a common failure. Details are:
>
> ISP = Demon
> Router = Various versions of Vigor (2600plus/i : 2600plus : 2600V :
> 2800) VPN = LAN-to-LAN, configured for "always on", connecting to
> another Vigor router
>
> About once every 6 weeks the router loses its internet connection.
> Inspection of the router (necessitating a site visit) shows that the
> ADSL sync is good, and reasonable figures are shown for up/down speed,
> SNR margin and attenuation. But the WAN status shows the PPPoA link is
> down. Clicking the button to "Dial PPPoA" has no effect.
>
> Rebooting the router, or unplugging and reconnecting the phone line,
> will always restore the conection. However the problem (by Sod's law)
> seems only to occur then the sites are unattended (as in over a
> weekend).
>
> On the one occasion that I struggled to get through to Demon's technical
> support they appeared to suggest that the router was lying, in that
> their system showed no ADSL sync; but I've little confidence in that.
>
> I have similar sites where the only difference is the ISP but these
> sites probably make less use of the VPN. I don't have any Demon sites
> where there is no use of any VPN, so can't speculate as to whether the
> VPN is a contributory factor.
>
> Has anybody seen anything similar? If so, have you a resolution?
>
> Thanks,

I had similar problems for a couple of years with a Vigor 2800VG. About
a year ago Draytek released a firmware update claimed to give better
performance over poor lines. Vigor 2800 firmware 2.8.1 or later is what
you need. Since upgrading I have not seen the frustrating "line sync OK
but LPC timeout" problem. It worked even better for a colleague with a
truly pathetic ADSL line. Line length here is 4km, delivering about 3
mbit/s sync speed at 12.0-13.0 dB SNR.

Regards,

Mike