Print

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

[roseway]

[Don't upgrade to this version yet if you use a Thomson/Technicolor modem/router]

Changes since v4.50.1:

- v4.52.1 fixes "List index (-1) out of bounds" error with some setups

- while recording is in process the login details can't be changed (you now have to pause or stop recording first)
- fixed issue with invisible graph labels when using a dark desktop theme
- fixed incorrect detection of VDSL2 band plan with some modem/routers
- the alerts system can now send alerts to the event log as well as to email
- added more triggers to the alerts system (CRC, FEC, ES, SNRM return after low value)
- the email alert "From" address is now set to a sensible format
- now provides some limited support for D-Link DSL-3780 (more work to be done)

http://dslstats.plainroad.me.uk

[broadstairs]

Installed and running here. The alerts for CRCs to the Event log is already working well for me this morning.

Stuart

[roseway]

Thanks Stuart.

[kitz]

Im not sure whats gone wrong.  Ive tried redownloading again..  and Ive also rebooted the PC

Typical I didnt keep a copy of the old version, have you got a link so I can roll back please.


Out of bounds error.

[roseway]

Sorry about that. I'm a bit puzzled at the moment.

http://www.s446074245.websitehome.co.uk/files/dslstats32W-4.50.2.zip

[kitz]

Thank you eric, 
I'll try do some more testing to see if I can recreate exactly what happened, but atm it appears to be if I try to keep my .ini
If I start from fresh without using my old .ini it seems ok.


-----
ETA

Do I do something wrong when I upgrade to a latest version. 
Normally I just take the .exe file and let it overwrite the old one. Should I be doing something more?

[roseway]

Do I do something wrong when I upgrade to a latest version. 
Normally I just take the .exe file and let it overwrite the old one. Should I be doing something more?

The file routers.dat changed between the versions, and it's possible that using the previous version with the new version of DSLstats would cause the error message you saw, but I'm not certain about that. I'll check it out tomorrow.

(Memo to self - in future, mention these changes when they occur with an upgrade.)

[burakkucat]

This latest release has been in use at The Cattery for around an hour . . .

So far there is nothing abnormal to report. 

[broadstairs]

Eric the alerts seemed to have stopped after midnight. This morning I can see conditions which should trigger an alert to the event log but it did not do it. I am not emailing alerts just recording to the log.

Stuart

[roseway]

Could you give me the specific details please, Stuart? i.e. What settings you have for alerts, and the relevant entries in the event log. Thanks.

[roseway]

I've updated the downloads to v4.52.1 which fixes the "List index (-1) out of bounds" issue with some systems.

[broadstairs]

Eric I have set for CRCs over 15 and FECs over 10000 only. They worked fine yesterday but the last log entry was for just before midnight. Since then I have had a couple of times when CRCs went over 15 but no log entries. The last log said

18 April 2014 23:50:40     Alert: Downstream FEC/min rose to 24415.86

after that no more alerts. I tried ticking and unticking the alerts but no change, Ive had 2xCRCs over 15 and 2xFECs over 10000 but no log entries. Rest of the log is working fine with all expected messages, just no alerts.

Stuart

[roseway]

Stuart, it's working as designed (although you may take the view that the design needs changing ). What happens when an alert is triggered is that the value which caused the alert becomes the new trigger threshold. The trigger thresholds are reset to the values you entered if you stop recording, there is a re-sync, or you edit one of the values (which resets just that value).

I did it this way to avoid the possibility of an endless stream of alerts if the line becomes very noisy while you're not present to notice it. It would be simple enough to change this behaviour, or make it optional perhaps.

[broadstairs]

Stuart, it's working as designed (although you may take the view that the design needs changing ). What happens when an alert is triggered is that the value which caused the alert becomes the new trigger threshold. The trigger thresholds are reset to the values you entered if you stop recording, there is a re-sync, or you edit one of the values (which resets just that value).

I did it this way to avoid the possibility of an endless stream of alerts if the line becomes very noisy while you're not present to notice it. It would be simple enough to change this behaviour, or make it optional perhaps.

I can see your point however in the case I'm looking for I set the initial threshold higher than any normal value for for both thresholds and I do need to keep that initial value so perhaps setting it as an option is the best way forward. When looking for spikes it is the only way to do it.

Stuart

[roseway]

I can see your point however in the case I'm looking for I set the initial threshold higher than any normal value for for both thresholds and I do need to keep that initial value so perhaps setting it as an option is the best way forward. When looking for spikes it is the only way to do it.

And in turn, I can see your point. I'll make it optional next time.