Kitz Forum
Broadband Related => Router Monitoring Software => Topic started by: roseway on September 01, 2013, 03:51:59 PM
-
This is mainly a fix for the "turn of the month" problem discussed below, but there are also a couple of other small changes.
Changes since v3.92:
- corrected issue which caused program to crash at the start of a new month
- sample timing is now linked directly to the system time, so it shouldn't drift over time
- added option to save the event log automatically after each new event is logged
http://dslstats.plainroad.me.uk
-
Its up and running now... looks good and since it is day 1 of month I am not really bothered about collecting the missing data, I've restarted it using my saved .ini files etc and all seems OK. In fact it looks to me like it has caught up with traffic totals pretty well.
Stuart
-
Thanks for confirming that, Stuart.
-
It's probably down to my own doing but I see something odd . . . ???
A screen-scrape is attached.
-
Excuse my thick-headedness, but what's odd?
-
:doh: D'oh!
It's b*cat's failure to read the graph's legend correctly that is 'odd'! I put it down to a lack of food in my bowl . . .
Please ignore my caterwauling. :-[
-
Have a tin of pilchards on me. ;D
-
A kind gesture, thank you but . . . :sick: . . . I don't like fish! :angel:
-
Eric something is not right with the traffic collection. This morning my daily graph shows no data for day 1 at all it is only showing day 2. The monthly graph also shows nothing for either August or September. I restarted tis latest version using the original files from the failing version sadly though I just renamed them not copied them so whatever is there now is only what this new version has done with them.
Stuart
-
Stuart, I think this is probably the result of yesterday's events, rather than something fundamentally wrong in v3.94. The traffic monitoring hasn't changed between the previous version and this. The best option, I believe, would be to reset the monthly traffic and accept that some earlier data is lost. I'm sorry about that.
-
OK but that does not really explain why the monthly graph shows nothing for September and the daily graph only shows data for today. I'll do as you suggest and see what happens tomorrow.
Stuart
Update: Just twigged why the month graph was not showing anything for September, I had set Monthly Limit which shows on the graph and at 40GB means that there was so little to show for September at present it looked like nothing was showing. Not sure I like the idea of showing the limit on graph as on by default, maybe an option would be better or perhaps only turn it on when 75% of month limit has been reached.
-
Just thought I'd compare my traffic from DSLStats with what TT actually believe is my download traffic and while there is a difference it does put both values in the right ball park. The differences can be explained by my DSLStats data started on day 2 this month, also TT's data is a couple of days behind actual so some calculation is needed.
So at least from my perspective DSLStats gives me a good approximation of my actual traffic tracked by TT.
Stuart
-
Thanks for that information,Stuart. To be honest, I'm getting more than a little frustrated at the inconsistency and variability of the traffic data reported by routers. I've a few reports now of the same router having different interfaces in different situations (the HG622 being one such example). I've now switched back to using the HG612, and I find that the interface which delivered the correct data the last time I used it (atm1) now fails to deliver, and the ppp256 interface gives proper information. The value reported by atm1 has stuck at 4294967295 bytes (2^32 - 1).
I don't think I'm going to spend any more time on this feature, as it seems to work properly once you find the appropriate interface, and there's little or nothing I can do to automate that. I'm going to tidy up the instructions on my website, and then get on with something more rewarding.
-
I think that is perfectly reasonable. I was hoping to get some idea from it of my traffic and that's exactly what I'm getting.
stuart
-
I appreciate your comments, Stuart.
-
Thanks Eric for mentioning the HG622. I used to have the traffic working for the HG622 on ADSL2+ but having just gone back to the HG622 I can't find anything that works or even corresponds to what telnet ifconfig gives for any inferface. It would have been an earlier version of DSLstats last time (I am currently on 3.82 -too lazy to change when all else seemed to be working OK). I will try 3.94 and see it helps but I am not expecting anything.
-
My HG622 works on ADSL2+ using ifconfig with 0.0.38
Stuart
-
My HG622 works on ADSL2+ using ifconfig with 0.0.38
After a lot of experimentation, I agree with that configuration. :)
-
Thanks both. Unfortunately, the last time I fired up my HG622, there was no 0.0.38 interface, and the interface which reported traffic correctly was called nas_0_38. ???
-
I really don't think you can do any more with that feature. :)
To be honest, when you mentioned that you would attempt to implement it, I thought it might prove to be a 'feature too far'. :-\
-
That's my conclusion too. As it happens, this is one feature which should work with no problems with Technicolor routers, because the custom CLI in those routers provides the information in an unambiguous form.
-
Well... fresh power up and moved to 3.94. Traffic currently working on the HG622 OK on 0.0.38 - it did not before - and from Eric's comments maybe it won't one day.
I notice that the ES seems to start counting from router power up and not DSLstats startup. It also seems to not allow for the initial short averaging period giving me fright of 300/hour at the start of the first hour dropping to a usual 2 or 3 /hour after an hour or so. I don't recall this happening before but maybe it did? - not a big issue if you expect it.
Also one seemingly endless loop of error exceptions occurred and DSLstats had to be killed with the task manager (Windows XP). I have not been able to reproduce that so far, so it may be an initial oddity.
-
I'm currently doing a bit of refining of the ES section, so I'll check that out.
The rash of error exceptions is a mystery at the moment. I hope it doesn't happen again, but I will see if I can think how it might have happened.