This is exactly why none of my friends or family have ST yet. I know probably 10 people that would put it in tomorrow if I gave it the thumbs up, they all love my setup. But none of them would be able to deal with all the issues.
My Sunrise/Sunset finally started working well again last week. Now since monday nada. Also had two of my random devices just stop working again. A hub reboot solved that. No way I could recommend it at this stage to early in the beta. One day it will be out of beta and stable enough for the average Joe.
I’ve had about a dozen failures between last night and this morning. Even updating the RM rules isn’t bringing my porch lights back. I’m patiently waiting out the new scheduler, and just periodically go into the app to make sure my bathroom fan hasn’t been on for an hour…
This is admittedly confusing, but just to clarify - make sure you are clicking “Update” on the specific instance of Smart Lighting (or any other app) that is causing issues.
For example, if my automation named “Turn off Centralite Switch at 9:30 AM and off at 6:30 PM” is having issues, make sure to click Update next to it - not on the “Smart Lighting” parent.
Maybe this is already clear to everyone, and in that case, carry on
I setup RM to one button update all my motion delayed rules, hit that last night, and manually hit update on all my sunrise/sunset rules then made sure all the lights where on the proper setting before going to bed. In the morning, I had either a 10+ second delay on motion, no auto off, and the porch light was still on way past the scheduled time.
Not complaining, just stating what’s been happening. I’m anxiously awaiting the new scheduler, but it concerns me the update button seems to not be working for me either. My bathroom fan DID turn off late this morning, at 23 minutes vs the 20 it’s supposed to, so the problem is not occurring ALL the time.
No doubt it’s confusing. Calling it an “update” is confusing in the first place. First thing that comes to mind is that the app is out of date and needs updating.
All my rules were broken as well, and I was able to open some of them and clikc Done to fix. Though many of mine are now giving unexpected errors. I’ll have to try the update rule.
Actually it looks like all my rules are basically broken today, in terms of being able to open them. Weird.
I have a rule with a very odd situation. It is a delayed motion detector. Pretty typical I think. Anyways, it has been acting all wacky, so I had to “Update” it. Then it showed no scheduled jobs. But if I hit Refresh from the Rule details in the IDE (upper right link) then an old job shows up. Is there any way to clear this short of removing the rule and recreating it?
This is just my hypothesis buy i believe that crap fails because their servers are overloaded and once a schedule is missed it is toast. Now that the have given us the handy update button which incidentally is not working for me, it appears as if they are purposely having events fail to aleve the server overload, and then have support tell us use this neat new feature. Why else has everything gotten much worse since it was released?
Why aren’t these issues being made public to the community through a pinned announcement? I’ve convinced several family members and a co-worker to buy Smartthings and we all had to search the forums to see that is a known issue…
Maybe Samsung can give these guys an increase in their allowance or something so Smartthings can buy some upgrades from Newegg?
Seriously though, I was excited when Samsung announced such a big investment… But it seems to have all gone towards converting to Tizen and future products?
What? Who said we are converting to Tizen? That’s just weird. [quote=“Lgkahn, post:72, topic:38054”]
This is just my hypothesis buy i believe that crap fails because their servers are overloaded and once a schedule is missed it is toast.
[/quote]
Yes it is a scaling issue. One we hope to have resolved soon. The guy working on this is very impressive.
The GUY working on this, like 1 guy, are you kidding? A complete failure of your timing engine and you have 1 guy working on it!
For a scaling issue, caused by adding more users, caused by selling the hubs that you manufactured and sold, seriously, how much of a surprise could it have possibly been when those hubs started using the scheduler? I really think somebody at SmartThings needs to take a serious looks at their development team and there testing procedures because both seem to be not working.
Poor, very very poor, if i treated my customers this way i would very rapidly not have them any more.
Moving to Tizen was widely reported (speculated rather) after the initial Samsung announcement. Glad to hear that it isn’t true, seemed like an unnecessary reinvention of the wheel. I admit only following this stuff closely when I want to tinker or things stop working.
Right now I’m hearing from everyone that scheduler problems have exponentially gotten worse over the past few weeks. I’m wondering how much of that was compounded by the new “update” feature and RM being introduced, as well as RM now being automated to update as well.
I’m starting to think that when RM came out and started to use more and more delay with cancel, and time based rules that all of the sudden people started to put a much heavier load on scheduler. Then in an effort to alleviate some of the newfound issue the ability to hit update drove that much more load onto the system because now in addition to the normal scheduler loads, everyone hitting update added another level of load. Then as Bruce automated the ability to automate all rules with an update it drove more scheduler cycles…
Hence the almost collapse of scheduler under the newfound load created by the various implementations to work around it. Good intentions but instead a near death blow as cycles exponentially increased…? Could be way off but it seems odd that it got that much worse as the tools were made available to utilize it more…
There’s just one small problem with your theory: Only a tiny fraction of ST users ever come on this Community, so only a tiny fraction ever see the Update feature, either in the IDE or Rule Machine. Only a tiny fraction of ST users are using Rule Machine for the same reason, as it is only available within the Community. So for your theory to be right, that tiny fraction of users have degraded the performance of the ST cloud and scheduling. Possible, but if true, then ST’s problems are considerably more severe than we think.
The more likely explanation is the overall growth of the ST user base, the vast majority of whom never use Update or Rule Machine.
One day it will be out of beta and stable enough for the average Joe.
