time Time Drift - feedback

[limelight 0]

HC2 with 4.143.

Uptime 7h. Time 7mins fast

Uptime 14h. Time 8 minutes fast.   

 

 

Hcl with 4.143

uptime 20h time 6 mins slow

 

 

 

[samuel 27]

Is there not a trigger, so that the NTP will update time - date on the HC2/HCL on a daily basis, when online and connected to LAN/Internet?

Edited November 21, 2017 by samuel

[petergebruers 1,268]

17 minutes ago, samuel said:

Is there not a trigger, so that the NTP will update time - date on the HC2/HCL on a daily basis, when online and connected to LAN/Internet?

 

In this rather long post, I motivate why time adjustments should be continuous, and not done "once every x hours" or so. This is my personal view, feel free to disagree and discuss this with me after reading it...

 

 

[samuel 27]

it should be done at least ones a day an update so that time/date will be updated as long as there is an active internet connection.

[T.Konopka 8]

Guys, thanks for the continuous feedback.

[sonnyboy 11]

3 min i  3 days with 4.144 Beta.

Hc2

[chapa 3]

4.160

3 minutes going forward, for 7 days running so far.

 

Any idea how this problem can be resolved?

[samuel 27]

V4.160 running already for 8 days, with 2second timedrift

Please login or register to see this attachment.

Please login or register to see this attachment.

Edited February 1, 2018 by samuel

[petergebruers 1,268]

27 minutes ago, chapa said:

4.160

3 minutes going forward, for 7 days running so far.

 

Any idea how this problem can be resolved?

 

This is a long shot... Set clock to manual and adjust... reboot.... check if time is OK... set back to "Update from internet" ... Reboot...

 

I'm mentioning this because I do not have a large time drift (few minutes per month) and when the time difference annoys me, I follow that procedure. I cannot remember when I last did that, probably a month ago so while running one of the 4.15X betas. I install every beta, I do frequent inclusions so lots of reboots, but I do not schedule reboots at regular intervals.

 

Now @samuel mentioned his clock is pretty good, I decided to check mine... It is pretty much "spot on" ( < 1 sec deviation from other clocks) after 49 hours uptime. I'm surprised it is that good. I'm using "be.pool.ntp.org".

 

This might be a coincidence... I do not know... Maybe they did change something.

[chapa 3]

thanks @petergebruers

 

just did your suggestions.

will report in day or two. will be enough time to observe if the trick seems to work

[Igy 23]

I dont know why, but in my case, timedrift is gone. Uptime 18days and no timedrift. Before I was several minutes off after such a long uptime.

[cag014 730]

+1

[Thomasn 77]

11 days of running. No timedrift!

[T.Konopka 8]

Good to know!

[chapa 3]

For 25 days, Fibaro time went almost 4 minutes behind.

It is much better than 3 minutes forward just for a week (as used to be before 4.160 and following @petergebruers instruction).

 

Please login or register to see this image.

/monthly_2018_03/image.png.668973348c76da760b45c6545a0d64ac.png" />

 

Unfortunately, it seems the problem is still there and need attention.

 

Edited March 12, 2018 by chapa

[Hiram K. Hackenbacker 2]

How can it be so difficult to keep the system time running ON time? Doesn't  NTP work in Poland? When I discovered this time drift, I've written a script (took me an hour or so) that syncs with my Axis camera (they hired real software engineers, obviously) in my home network when the deviation gets too big. This was the quickest option.

 

While doing that, I found that setting the time via the API  causes a reboot and a NTP-sync of the clock (hurray; the Fibaro programmers managed to do that!). So my script calls HC2 on the local host with a time setting of 00h00 (t doesn't matter what time value is entered), the HC2 reboots (no clue why this is necessary for a time change) and runs in sync with the rest of the world for a couple of hours....

 

I also found that setting the time manually through the interface has the same effect (finally some consistent behaviour), but what is the use of a manual setting if it syncs the system with NTP? And if I select "Download time from the Internet", it runs out of sync after a day or so?!?

 

I do not use words like the following easily, especially not on a forum, but I think the Fibaro programmers are a bunch of n00bs (believe me, I am tempted to use some stronger terms here). I am a former system programmer myself and upon finding bugs like this in my own software, I would not have slept nor eaten until they were properly solved! IMO there are two types of software bugs: excusable and inexcusable. The first type is caused by anything ranging from: "Software can be too complex for humans to grasp" or "The programmer had too much to drink the night before and goofed up, after all he's just human". The second type of bug (inexcusable) is caused by anything ranging from "The programmer is totally incompentent and should start with writing test programs to get his skills right" to "The programmer doesn't give a toss about the quality of his work and leaves at 5 o' clock each day to get home in time". So far the bugs I've seen in the Fibaro system are of type two, i.e. inexcusable.

Edited April 29, 2018 by Hiram K. Hackenbacker

[petergebruers 1,268]

39 minutes ago, Hiram K. Hackenbacker said:

but I think the Fibaro programmers are a bunch of n00bs

 

Did anyone call you "n00b" lately? And did that motivate you, to work better/harder?

[Hiram K. Hackenbacker 2]

No, but if someone did, I would certainly draw conclusions or, at least, evaluate my own behaviour and skills. As I state above, I am reluctant to use terms like this because, as you say, some people are not motivated by words like that. But the level of incompetence and/or the unwillingness to fix bugs that I see with Fibaro is really shocking! I'll give another example of this: the Variables Pane shows variable values in fields that are too small to even show the full value of a long integer, whereas there's plenty of space. A simple, but annoying bug like this can be solved by a compentent programmer in minutes! Then, why is it still there? With the latest release (4.180) they solved only ONE issue! Why not this quick win? By the way, I've already heard that with 4.180 another, very serious new bug was introduced...

[petergebruers 1,268]

@Hiram K. Hackenbacker thank you for sharing some facts and your opinion, I think this works a lot better than calling someone an "N00b". I understand your frustration, I have been using this product since 2013. A lot of progress has been made, but issues remain... I hope you can see the positive elements too, and use this Fibaro community to learn and share solutions as well... I do not know you (yet) and you only have a few posts so I might tell you things you already know. For instance, I  recommend you to evaluate another controller software (either open source or buy something like Z-Way, Homeseer, Vera, whatever) or at least read some posts on their forum. Just to put things in perfective. In any case: have fun with home automation!

[Hiram K. Hackenbacker 2]

Thanks for your advise about trying alternatives! 

My experience with Fibaro is relatively short, but I know people that have a lot longer experience and it is sad to say that they are not very positive either about the quality of the software. Especially the fact that each new release solves very little and introduces new, nasty bugs each time. The issue here is that Fibaro is not cheap, so one may have high expectations. I got my HC2 for free from a friend, and it gave me the opportunitiy to polish (pun not intended) my rusty programmers' skills and build a nice fully automatic alarm system (so my overall experience is positive), but I can tell you that if I had paid the 500+ euro's that it costs over here, I'd sent it back to the shop and asked for a full refund in not very nice wordings, I can assure you.

 

By the way, I hope that you see that I've used the term "n00bs" with ample arguments; I totally agree that a forum is no place to call people names, but in this case they (Fibaro) leave no other option. The most disturbing fact is not that there are bugs (everybody makes mistakes and "software is a bitch"), but the perceived unwillingness or incompetence of Fibaro to solve them and to improve their product.