Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[burakkucat]

You will see from the attached montgae that there is some data missing for a few days.  BE confesses he was messing with the scripts to try to log & graph more data but apparently stuffed it up !!!  Hence my great interest in your new ones 

Oops!

Typical of Eagles!     

[Bald_Eagle1]

Thanks Paul, will email you if any issues 

Have you had a dabble with the updated scripts yet?

I'm still not 100% happy with the data harvesting error checking/correction yet.
I think I have an issue at around 05:00 am, where a few minutes of harvested data are "missing" (only some mornings though), possibly when AVG is either virus checking or updating its database.

This does seem to be a Windows 7 issue only, but I haven't yet sussed whether AVG, Windows 7, the scripts, or "something else" is/are to blame.

[Blackeagle]

Ha, yes I have.  Attached is a montage 

I note with interest that early this morning, DLM resynced the line with a higher downstream sync, this after 72 hrs of stability.  It'll likely go back down again, but at least DLM is moving it up now, whereas before BTOR reset it, it was only going in a downward direction.

The only data missing from the graphs AFAIK is a slightly greater than 2hr period yesterday evening, when due to a technical issue, windows would not boot   

-=EDIT=-

Although the downstream errored seconds graph reports a figure in the heading, this does not seem to be graphed against time ?? 

[Bald_Eagle1]

-=EDIT=-

Although the downstream errored seconds graph reports a figure in the heading, this does not seem to be graphed against time ?? 

Ah, I see you have also spotted that "glitch".
I have already emailed you to mention it & to request some log files for me to investigate this.

At any given time, the error count plots should show 60 seconds or less & the total Error Seconds in the graphed period should be only the number of error seconds during whatever period you choose to plot in days, hours or minutes.

The Y axis adjusts itself to the maximum No. of error seconds seen in any 1 minute period.
It should usually max out at only a few, as per my 24 hour example attached - 96 DS error seconds within the 24 hour period.

EDIT:

Also, the US Error Seconds Y axis should grow or shrink to match, just to keep them equal for quick/easy comparison purposes.
The Y axis for the DS & US Interleaving graphs should not match each other though.
This is because some connections have DS Interleaving depths in the thousands, yet US Interleaving can be 1 (OFF) or some other very low value.

[Bald_Eagle1]

The Full Monty montage is not ordered & is put together with this command:-

montage *.png -quality 100 -geometry 640x -border 5 -tile 3x line_stats_FULL_MONTY_P-%FILEDATE%.png

To create the montage with a bit of order to it, you could replace that with:-

Code: [Select]

montage snrm.png CRC_errors.png FEC_errors.png sync.png DS_RSUnCorr.png^
 OHFErr.png attainable_rate.png HEC_errors.png DS_RSCorr.png power.png DS_RS.png US_RS.png DSL_UP_TIME.png DS_INTERLEAVING.png US_INTERLEAVING.png^
 DS_ES.png US_ES.png DS_OHF.png DS_line_atten.png US_line_atten.png US_OHF.png DS_signal_atten.png US_signal_atten.png^
 US_RSCorr.png US_RSUnCorr.png^
 -quality 100 -geometry 640x -border 5 -tile 3x line_stats__FULL_MONTY_P-%FILEDATE%.png


[Blackeagle]

OK, I have edited graphpd.bat to update the full monty montage and have attached two examples.

The first is a 24 hr period (covering the last 24 hours) and shows everything correctly, the second is a two day period (which includes the broken data and covers the last 48 hrs) which has the same US/DS error errors ?!? as we have already discussed !! 

I presume as long as I do not include that iffy data that everything will continue to graph correctly ?

BTW, I do like the improved folder naming where it specifies the range covered.  Nice touch Paul 

[Bald_Eagle1]

OK, I have edited graphpd.bat to update the full monty montage and have attached two examples.

I presume as long as I do not include that iffy data that everything will continue to graph correctly ?

Yes, it only does that when it plays catch up from either a fresh start of the logging, or when there has been a pause in logging.
You could just delete the "offending" row(s) for it to be correct again.

BTW, I do like the improved folder naming where it specifies the range covered.  Nice touch Paul 

Ah yes, I got tired of manually adding the graphed range to the folder name for archive purposes & sometimes forgetting which made it difficult to trawl back to relevant graphs when needed.

[yorkie98]

[/quote]

Right,

Firstly the big red 'V' shape in your bit-loading & SNR graphs for the downstream D1 band looks very unusual.
It would be interesting to see the Attenuation & SNR values from --pbParams.

[/quote]

I know this is going back in time quite a bit but I think I have solved the mystery of the V shape in the graph.
It appears that for VDSL this IS normal. It is a deliberate filter applied to these frequency bands to avoid interference with ADSL connections in the ducts/cabinets.
There is more information here : http://www.openreach.co.uk/orpg/home/products/industryforums/superfastfibreaccess/downloads/FTTC_TWG_slide_deck_19th_Mar_2009_v4.pdf
In pages 29-32

HTH..

[Bald_Eagle1]

There is more information here : http://www.openreach.co.uk/orpg/home/products/industryforums/superfastfibreaccess/downloads/FTTC_TWG_slide_deck_19th_Mar_2009_v4.pdf
In pages 29-32

HTH..

I'm afraid that link appears to be broken, causing a runtime error at my end.

I would mention though that from our limited data, such a pronounced 'V' does now appear to be more prevalent on connections to ECI cabinet DSLAMS than to Huawei cabinet DSLAMS.

[yorkie98]

I just tried the link, still working here, must be a problem launching embeded Acrobat reader in your browser.
Try right-clicking the link, save as then run the file outside of the browser.
There are graphs on it showing the filtering applied and these very much mirror the frequency ranges of the V notch in the traces given.

[Bald_Eagle1]

It was an AVG Toolbar issue.
I have disabled it & hey presto, problem gone.

I will study the document later.

FWIW, my graphs look like the attached (Huawei modem on Huawei DSLAM).

[GigabitEthernet]

How is the connection now?

[NewtronStar]

It was an AVG Toolbar issue.
I have disabled it & hey presto, problem gone.

I will study the document later.

FWIW, my graphs look like the attached (Huawei modem on Huawei DSLAM).

Hello BE1 I have noticed your download & upload sync has gone up since june

How did you get the increase ?

My stats seem to be steady 

[Bald_Eagle1]

Hello BE1 I have noticed your download & upload sync has gone up since june

How did you get the increase ?

That was immediately following a DLM reset (didn't last though).

My connection is now completely stable, but at lower speeds with Interleaving on.

[NewtronStar]

Well at least you were able to try a DLM reset 

It looks like I am not going to get anymore out of my line, but still I am happy with the 27.5 Meg Down & 5.5 Meg Up from my 700-850 meters line length !