HVS-2 problems


Ernst Lobsiger
 

Dear All

Graham contacted me first re this issue. HVS-2 has an ongoing problem of missed/recovered packets
at very high levels. This is a problem seen all over Europe. That's why it's likely an Uplink Chain issue.

See it on David's site, scroll down to  -->  EUMETCast HVS-2  - lost TelliCast packets

http://satsignal.eu/mrtg/performance_eumetcast-europe_losses.php

Has anyone more information about this problem? Maybe someone (Graham?) should contact OPS.

Regards,
Ernst



David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

On 11/03/2021 19:56, Ernst Lobsiger via groups.io wrote:
Dear All
Graham contacted me first re this issue. HVS-2 has an ongoing problem of missed/recovered packets
at very high levels. This is a problem seen all over Europe. That's why it's likely an Uplink Chain issue.
See it on David's site, scroll down to  -->  EUMETCast HVS-2  - lost TelliCast packets
http://satsignal.eu/mrtg/performance_eumetcast-europe_losses.php <http://satsignal.eu/mrtg/performance_eumetcast-europe_losses.php>
Has anyone more information about this problem? Maybe someone (Graham?) should contact OPS.
Regards,
Ernst
Thanks, Graham & Ernst.

Yes! Reported to EUMETSAT this morning.

BTW: each service has its own URL, for example:


https://www.satsignal.eu/mrtg/performance_eumetcast-europe_losses.php#hvs2

Cheers,
David
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


Ernst Lobsiger
 

On Thu, Mar 11, 2021 at 10:43 PM, David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺 wrote:
Yes! Reported to EUMETSAT this morning.
David,

it seems the problem is linked to a traffic increase of more than 10%.
What new stuff has been added to HVS-2 yesterday around noon?

Cheers,
Ernst


David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

On 12/03/2021 16:10, Ernst Lobsiger via groups.io wrote:
David,
it seems the problem is linked to a traffic increase of more than 10%.
What new stuff has been added to HVS-2 yesterday around noon?
Cheers,
Ernst
I don't know, and I can't see anything obvious (for HVS-2) in the Manufacturers' Notes. May be in the weekly messages? I didn't get an immediate reply from Ops, though, saying you've missed (whatever).

Cheers,
David
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


Ernst Lobsiger
 

On Fri, Mar 12, 2021 at 11:13 AM, David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺 wrote:
I don't know, and I can't see anything obvious (for HVS-2) in the Manufacturers' Notes
David and All

When I  compare march 10 - 11 and 11 - 12 of March from 12:00 to 12:00 I see a 20% mean traffic
increase and an almost 10% maximum traffic increase. This increase is seen on satsignal as well:

http://www.satsignal.eu/mrtg/performance_eumetcast-europe_rx_io.php#t2

BEWARE: My plots are in Bytes/s while satsignal uses more common Bits/s.

Cheers,
Ernst


Ernst Lobsiger
 

All

Here is one more hint: TD-15 on page 11 says HVS-2 has a maximum 85 Mbps info bit rate and 77 Mbps IP data rate.
On satsignal Arne van Belle reaches the maximum Rx I/O of 85 Mbps while all others are only slightly below this limit.

My yellow graph is MultiCast traffic in Bytes/s. So this might eventually not be what EUMETSAT calls IP data rate.
IP data rate 77 Mbps could be what we would see in the clients WEB interface if we subscribe for just everything.

In any case if I multiply 10.4 by 8 I get far beyond these 77 Mbps. The yellow graph also shows signs of saturation
at 10.4 (flat top). Did EUMETSAT just overload the system? Any other ideas regarding these maximum bitrates?


Regards,
Ernst


David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

On 12/03/2021 19:28, Ernst Lobsiger via groups.io wrote:
David and All
When I  compare march 10 - 11 and 11 - 12 of March from 12:00 to 12:00 I see a 20% mean traffic
increase and an almost 10% maximum traffic increase. This increase is seen on satsignal as well:
http://www.satsignal.eu/mrtg/performance_eumetcast-europe_rx_io.php#t2 <http://www.satsignal.eu/mrtg/performance_eumetcast-europe_rx_io.php#t2>
BEWARE: My plots are in Bytes/s while satsignal uses more common Bits/s.
Cheers,
Ernst
Interesting that the reported difference between T1 and T2 link margin increased at about the same time:

https://www.satsignal.eu/mrtg/lund-snr-diff.html

Probably not a real difference, though, just an artefact of the increased traffic?

David
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

On 12/03/2021 20:03, Ernst Lobsiger via groups.io wrote:
All
Here is one more hint: TD-15 on page 11 says HVS-2 has a maximum 85 Mbps info bit rate and 77 Mbps IP data rate.
On satsignal Arne van Belle reaches the maximum Rx I/O of 85 Mbps while all others are only slightly below this limit.
My yellow graph is MultiCast traffic in Bytes/s. So this might eventually not be what EUMETSAT calls IP data rate.
IP data rate 77 Mbps could be what we would see in the clients WEB interface if we subscribe for just everything.
In any case if I multiply 10.4 by 8 I get far beyond these 77 Mbps. The yellow graph also shows signs of saturation
at 10.4 (flat top). Did EUMETSAT just overload the system? Any other ideas regarding these maximum bitrates?
Regards,
Ernst
Yes, it looks like it's just been pushed too far. I would hope that whatever change was made was documented at the time!

My understanding is that that the bits from the receiver include more than just the data, things like PIDs and other DVB-S2 control data. Hence the 77 Mbps read data from a maximum bit rate of 85 Mbps.

I did also look for bad DVB-S2 packets or frames, but saw none.

Cheers,
David
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


Ernst Lobsiger
 

On Sat, Mar 13, 2021 at 01:49 AM, David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺 wrote:
My understanding is that that the bits from the receiver include more than just the data, things like PIDs and other DVB-S2 control data. Hence the 77 Mbps read data from a maximum bit rate of 85 Mbps.
David,

YES and there is redundancy because of the TelliCast FEC on the UDP packet level (missed/recovered). If I multiply my 10.4 with 8 bits per Byte I get slightly below 85Mbps like everybody else.
BTW: We have already seen a net increase of the HVS-2 traffic on February 25th 2021. This is shown in my eLuna graphs but can be seen in all MRTG Rx-I/O monthly traffic graphs on your site.

Regards,
Ernst




Ernst Lobsiger
 

Dear All

Starting today 14:45 UTC the HVS-2 missed/recovered packet storm has been noticeably reduced.
It seem the max traffic has been very slightly reduced as well to about 82 Mbps (rather hard to say).

On my receivers March 12th and 15th I have absolutely perfect HVS-2 TCLogSummaries while the
missed/recovered files are catastrophic to say the least. The multicast interface looks perfect too.

This made me think that the problem could as well be with our local HVS-2 tc-clients that just show
nonsense on the WEB interface regarding missed/recovered packets when traffic goes to 85Mbps.

Any other ideas or feedback from OPS?

Regards,
Ernst


David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

On 16/03/2021 18:51, Ernst Lobsiger via groups.io wrote:
Dear All
Starting today 14:45 UTC the HVS-2 missed/recovered packet storm has been noticeably reduced.
It seem the max traffic has been very slightly reduced as well to about 82 Mbps (rather hard to say).
On my receivers March 12th and 15th I have absolutely perfect HVS-2 TCLogSummaries while the
missed/recovered files are catastrophic to say the least. The multicast interface looks perfect too.
This made me think that the problem could as well be with our local HVS-2 tc-clients that just show
nonsense on the WEB interface regarding missed/recovered packets when traffic goes to 85Mbps.
Any other ideas or feedback from OPS?
Regards,
Ernst
Nothing yet, Ernst. Could be the virus is delaying a response. Agreed that zero file losses are "strange", and I'm seeing the same here (high winds and rain permitting). I'd be surprised if it's the Web interface, although that does seem possible.

Data rate exceeds specification gives errors - problem explained. But we still don't know what changed - it certainly wasn't everyone's client and at least three different receiver types are showing the effect. Harmut Schulla's graphs since ~midnight are interesting!

Cheers,
David
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


Graham Woolf
 

Hi All

I have just heard from Andreea at OPS Eumetsat that the HVS-2 issue has now been fixed

My HVS-2 reception is OK this morning

Regards

Graham


James Brown
 

Yes indeed from yesterday afternoon. 

James

On 17 Mar 2021, at 10:02, Graham Woolf <woolfg@...> wrote:

Hi All

I have just heard from Andreea at OPS Eumetsat that the HVS-2 issue has now been fixed

My HVS-2 reception is OK this morning

Regards

Graham


Ernst Lobsiger
 

On Wed, Mar 17, 2021 at 03:02 AM, Graham Woolf wrote:
I have just heard from Andreea at OPS Eumetsat that the HVS-2 issue has now been fixed

My HVS-2 reception is OK this morning
Graham and All

It seems the problem is not yet fixed 100%. Here is my HVS-2 graph compared witih Thorsten's HVS-2 missed/recovered packets timeline.

Cheers,
Ernst


Graham Woolf
 

Hi Ernst

Yes you are right mine is the same

You can also see my stats on Davids site - its Ringwwood

Regards

Graham


David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

On 21/03/2021 11:40, Ernst Lobsiger via groups.io wrote:
Graham and All
It seems the problem is not yet fixed 100%. Here is my HVS-2 graph compared witih Thorsten's HVS-2 missed/recovered packets timeline.
Cheers,
Ernst
Same here, Ernst. I reported that to EUMETSAT first thing this morning.

I noted that the interval between the losses is about 1 Sentinel orbit period.

Cheers,
David
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


Ernst Lobsiger
 

On Sun, Mar 21, 2021 at 04:52 AM, David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺 wrote:
I noted that the interval between the losses is about 1 Sentinel orbit period.
David,

it even seems that the missed/recovered bursts are in sync with strong multicast traffic drops.

Regards,
Ernst


David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

On 21/03/2021 12:28, Ernst Lobsiger via groups.io wrote:
David,
it even seems that the missed/recovered bursts are in sync with strong multicast traffic drops.
Regards,
Ernst
Trying to push that extra data through when there /might/ be room!

David
--
SatSignal Software - Quality software for you
Web: https://www.satsignal.eu
Email: david-taylor@blueyonder.co.uk
Twitter: @gm8arv


Douglas Deans
 

On 21/03/2021 12:28, Ernst Lobsiger via groups.io wrote:
On Sun, Mar 21, 2021 at 04:52 AM, David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺 wrote:
I noted that the interval between the losses is about 1 Sentinel
orbit period.
David,
it even seems that the missed/recovered bursts are in sync with strong multicast traffic drops.
Regards,
Ernst
===========================================================================

I am not taking HVS-2 at the moment but I wonder if they are testing this :-

Dear Manufacturer,

The Sentinel-3A/B OLCI Level 2 Full Resolution (FR) NRT product is currently distributed on EUMETCast Terrestrial.

On 7 April 2021, we will start a trial dissemination of this product on EUMETCast Europe, High Volume Service Transponder 2:
- For Sentinel-3A OLCI L2 FR:

Channel: E2H-S3A-10

Multicast address: 224.232.224.12

PID: 610
- For Sentinel-3B OLCI L2 FR:

Channel: E2H-S3B-10
Multicast address: 224.232.224.13
PID: 610
Example filenames:

S3A_OL_2_WFR____20210317T095941_20210317T100241_20210317T115127_0179_069_307_3600_MAR_O_NR_003.SEN3.tar

S3B_OL_2_WFR____20210317T154915_20210317T155215_20210317T180358_0179_050_168_2700_MAR_O_NR_003.SEN3.tar

NOTE:
In order to receive the new channels’ data on dedicated target directories, users need to edit cast_client-channels_hvs-2.ini and add the following sections for the new channels:
[channel]

name=E2H-S3A-10

target_directory=<target dir…>

tmp_directory=<tmp dir…>

[channel]

name=E2H-S3B-10

target_directory=<target dir…>

tmp_directory=<tmp dir…>

If the sections above are not added to cast_client-channels_hvs-2.ini, the data will be received under the target directory of the default channel (name=*) in case it is defined.


Regards,
Douglas.


David J Taylor GM8ARV 🏴󠁧󠁢󠁳󠁣󠁴󠁿 🇪🇺
 

On 21/03/2021 15:21, Douglas Deans via groups.io wrote:
I am not taking HVS-2 at the moment but I wonder if they are testing this :-
Dear Manufacturer,
The Sentinel-3A/B OLCI Level 2 Full Resolution (FR) NRT product is
currently distributed on EUMETCast Terrestrial.
On 7 April 2021, we will start a trial dissemination of this product on
EUMETCast Europe, High Volume Service Transponder 2:
- For Sentinel-3A OLCI L2 FR:
Douglas,

Yes, that's possible, and when I first saw that I wondered where they were going to fit it in. I was unsure of the data volumes involved especially when "full resolution" was mentioned, but I think it's water temperature, so likely just a single channel. But tests before a trial are the norm.

Sometime this year Sentinel data is removed from HVS-2, IIRC, presumably to make way for MTG and EPS-NG tests?

73,
David GM8ARV
--
David
Web: http://www.satsignal.eu