larryllix Posted June 12, 2018 Posted June 12, 2018 I have been having a lot of problems with ecobee server going down lately. I believe they are trying to upgrade some of their server things, lately without telling anybody, of course. I would be nice for people to report that here so nobody starts to rebuild their NodeLink system. Many times the ecobee web access is also down and that indicates it is totally ecobee. Monday June 11, 2018 Today both my ecobee3 and 4 heartbeats stopped for hours. Later/tonight I noticed NodeLink popped up with a new pin, which I tried but ecobee refused every trial. Despite several tries to generate and new PIN NodeLink would not generate one?? I rebooted NodeLink, generated a new pin, and ecobee accepted it.
larryllix Posted June 19, 2018 Author Posted June 19, 2018 Monday June 18, 2018 Both ecobee3 and ecobee4 heartbeats stopped from aprox. 4:55 to 9:55 PM EDT. No NodeLink intervention required. Ecobee website access not checked.
larryllix Posted July 10, 2018 Author Posted July 10, 2018 Monday July 9, 2018 2018-07-09 22:23:15 NodeLink reports comm failures from ecobee. Ecobee web app will not respond. Has not cleared yet. 2018-07-10 01:12:54 - Ecobee: Error refreshing tokens - if this continues try setting a new app pin [ecobee1] Last error logged. Self corrected and worked fine since. Note: Always on Mondays?
rccoleman Posted July 10, 2018 Posted July 10, 2018 I noticed this last night while I was in the app on my iPhone. Both Ecobee3s were online and working and both suddenly dropped offline. Network connectivity was fine at the thermostats, but they both reported 'Connected to ecobee.com: No'. They both came online in a couple of minutes.
larryllix Posted July 16, 2018 Author Posted July 16, 2018 Monday July 16, 2018 2018/07/16 2:14:38 PM Both ecobees stop reporting heartbeat. Yup...Monday again. 2018/07/16 17:30... ecobee heartbeats never resumed, Ecobee nodes hung. New PIN issued on NodeLink main tab page. NodeLink shows new pin issued. Doesn't not work in the ecobee API website. Reports invalid PIN every time. NodeLink will not generate a new PIN same as last time. NodeLink ecobee3/4 nodes are hung again. Venstar node heartbeat was never reported failed. Rebooted Nodelink, displayed new PIN request, installed with ecobee and began functioning again. Sorry. This is all I logged with Level = Error. 2018-07-16 13:44:37 - Ecobee Web StatusError - ServiceUnavailable [ecobee1] 2018-07-16 13:47:38 - Ecobee Web StatusError - ServiceUnavailable [ecobee1] 2018-07-16 13:47:38 - Ecobee Timer Error - Expecting element 'root' from namespace ''.. Encountered 'None' with name '', namespace ''. [ecobee1]
rccoleman Posted July 16, 2018 Posted July 16, 2018 I saw this outage as well, and monitored the situation through all of the complaints on their Twitter feed. None of the outages have caused a problem with NodeLink, though. It seems to just pick up where it left off and recovers gracefully. I didn't need to do anything after the last outage and it's working fine.
larryllix Posted July 19, 2018 Author Posted July 19, 2018 Thursday July 19, 2018 2018-07-19 01:29:10 - Ecobee Web StatusError - ServiceUnavailable [ecobee1] .... 2018-07-19 02:05:11 - Ecobee Web StatusError - ServiceUnavailable [ecobee1] = self cleared. Both ecobee nodes heartbeats stopped Note: I had contact with ecobee the previous afternoon and complained about their API after their setback program API PIN installation stopped NodeLink API PIN. Reported previous post.
larryllix Posted July 24, 2018 Author Posted July 24, 2018 Monday July 24, 2018 2018-07-24 06:34:52 - Ecobee Get Error - Timeout [ecobee1] 2018-07-24 06:38:08 - Ecobee Get Error - Timeout [ecobee1] . . 2018-07-24 09:06:05 - Ecobee Web StatusError - ServiceUnavailable [ecobee1] 2018-07-24 09:09:06 - Ecobee Web StatusError - ServiceUnavailable [ecobee1] Monday again.
larryllix Posted August 1, 2018 Author Posted August 1, 2018 Tuesday July 31, 2018 Approximately 06:05 AM ecobee comms and app-site became unavailable Approximately 06:55 AM ecobee app-site came back online but NodeLink issued a new ecobee PIN. This locked out further comms until I reset the app PIN in the ecobee web app-site. (Thank Gawd for heartbeats provided to produce a notice for me and I was home for LAN access!!)
rccoleman Posted August 1, 2018 Posted August 1, 2018 I’m curious about why these outages are causing ongoing connectivity problems for you. I see the periodic outages, but I never have to do anything with Nodelink to recover. Ecobee support in Nodelink (and Nodelink in general) is nearly maintenance-free for me. It just works (thanks, io_guy!).
larryllix Posted August 1, 2018 Author Posted August 1, 2018 40 minutes ago, rccoleman said: I’m curious about why these outages are causing ongoing connectivity problems for you. I see the periodic outages, but I never have to do anything with Nodelink to recover. Ecobee support in Nodelink (and Nodelink in general) is nearly maintenance-free for me. It just works (thanks, io_guy!). I assume ecobee is doing some maintenance almost every Monday. This time early Tuesday morning but it destroyed my app PIN setting again. NodeLink usually self-restores but under certain response AFAIK NodeLink responded (mostly?) properly by issuing a new PIN to start over again. This is backed up by holes in their own served charting that match NodeLink heartbeats missing at the same times. Also, I caught ecobee changing settings in my stat after a support contact about their weather forecast, that just turned out to be lack of proper weather locations. They finally responded that they won't touch my settings again, but I know they are messing around with things related. That is really scary for me, especially after the really moronic responses I got from their support, for every support contact in the last few years. Maybe there is a lesson in all this. Never contact ecobee support about technical issues. Maybe never tell ecobee their API sucks dishwater? Don't get me wrong... I still firmly believe ecobbe offers the best thermostat around. This is about the 8th unit I have installed from many brands. I do miss the outside temperature sensor (due to micro-climate here) but I do NOT miss the temperature wandering around and short cycling problems other stats do.
larryllix Posted August 4, 2018 Author Posted August 4, 2018 Friday August 3, 2018 Random failures throughout the day. Ecobee website app also very slow with many pages not accessible or extremely slow. I upgraded NodeLink v0.9.17 to v0.9.18 and power cycled the RPi 3. Asus router was also upgraded to new release of firmware and then power cycled. Saturday August 4, 2018 Appeared to clear up for a few hours but then started again this next day. Ecobee web app-site continues to not respond to certain page requests. IE 11 and Chrome were tried with the same results. NodeLink reports missing poll responses sporatically. Most responses are successful in order to barely avoid triggerring my heartbeat monitoing program notifications. The disturbances still continue this morning.
rccoleman Posted August 17, 2018 Posted August 17, 2018 Both of my Ecobees have been working fine for a while (at least through last night), but today I noticed that the nodes were no longer responding to commands. I rebooted the ISY and restarted NodeLink, and now I see that no data is showing up in NodeLink, either. Initially I got an authorization error like this: Quote 2018-08-16 20:12:53 - { "error": "authorization_expired", "error_description": "The authorization has expired waiting for user to authorize.", "error_uri": "https://tools.ietf.org/html/rfc6749#section-5.2" }(P) [dnecobee] So I requested a new token and that message went away, but I was still not getting any data. I then noticed this in the log: Quote 2018-08-16 20:12:53 - Ecobee Web StatusError - InternalServerError [dnecobee] Presumably that means that the Ecobee servers are borked again? I can access the ecobee website and control my thermostats from there, so it's not completely broken. Giant log file attached, if that's helpful. logfile_2018-08-16.txt Edit: Nevermind, it's working again. I thought that I waited long enough for it to clear itself up, but I guess not.
rccoleman Posted August 19, 2018 Posted August 19, 2018 It seemed like it was broken again last night. I noticed that my incoming geofence program triggered, but my Ecobees didn't switch to "home" as expected and I couldn't change their state in the admin console. The odd thing is that the last two times I've seen failures, I've been able to see and manipulate the thermostats from both the web app and the iPhone app. Only Nodelink seems to have trouble with them. @io_guy I'm happy to send you my log from yesterday if you think there may be something interesting in there.
larryllix Posted August 19, 2018 Author Posted August 19, 2018 3 hours ago, rccoleman said: It seemed like it was broken again last night. I noticed that my incoming geofence program triggered, but my Ecobees didn't switch to "home" as expected and I couldn't change their state in the admin console. The odd thing is that the last two times I've seen failures, I've been able to see and manipulate the thermostats from both the web app and the iPhone app. Only Nodelink seems to have trouble with them. @io_guy I'm happy to send you my log from yesterday if you think there may be something interesting in there. Have you updated your RPi setup lately? It seems some others may be having problems with Venstar stats as well. Possibly mono is having some problems in updates lately or maybe RPi drivers? My last few RPi updates have my sound drivers garbling out badly. I am seeing occasional Venstar stat errors I haven't seen for months now. I also thought it may be my router for a while but an upgrade to that firmware didn't help..
rccoleman Posted August 19, 2018 Posted August 19, 2018 I updated a few months ago and did again when I saw the problem last week and again last night. I am using the official mono repo and have mono has updated each time.
larryllix Posted August 19, 2018 Author Posted August 19, 2018 25 minutes ago, rccoleman said: I updated a few months ago and did again when I saw the problem last week and again last night. I am using the official mono repo and have mono has updated each time. Does your file /etc/apt/sources.list.d/raspi.list show the right & matching o/s version (eg "jessie")? IIRC My file never updated through many upgrades and updates and was maintaining the wrong version of mono. I had to edit it by hand to eliminate a lot of random errors after updating the o/s so many times and from Jessie to Stetch. https://www.mono-project.com/download/stable/#download-lin-raspbian
io_guy Posted August 20, 2018 Posted August 20, 2018 8 hours ago, rccoleman said: It seemed like it was broken again last night. I noticed that my incoming geofence program triggered, but my Ecobees didn't switch to "home" as expected and I couldn't change their state in the admin console. The odd thing is that the last two times I've seen failures, I've been able to see and manipulate the thermostats from both the web app and the iPhone app. Only Nodelink seems to have trouble with them. @io_guy I'm happy to send you my log from yesterday if you think there may be something interesting in there. Do you get an error when you try to change it in the admin console? Enable debug on the Ecobee and System Config tabs and try to change a setting from the admin console.
rccoleman Posted August 20, 2018 Posted August 20, 2018 2 hours ago, io_guy said: Do you get an error when you try to change it in the admin console? Enable debug on the Ecobee and System Config tabs and try to change a setting from the admin console. Here's the initial attempt to set both thermostats to "Home" (at least somewhere in here). The two things that stick out to me are the fact that the tokens had expired, and then it got another "internal server error". I always have all debug settings set to "Debug", so there's a lot of output (error_output.txt). When I tried to change it from the admin console, it's really hard for me to pick out what was going on. It's was around 7:40pm or so, and I'm including the output of the ISY error log and the NodeLink log from the same time. It's not happening now, so all I have are the logs from last night (error_output_nl.txt & error_output_isy.txt). I'll keep looking through it to try to find the actual "set" commands from the admin console. I don't see an obvious "internal server error" similar to the earlier error. error_output.txt error_output_isy.txt error_output_nl.txt
io_guy Posted August 20, 2018 Posted August 20, 2018 I see way more webrequests (from the ISY) than responses from the Ecobee. I'll add some additional logging to the next version to capture each send to the Ecobee (not just response) so they can be matched up to help debug this. Also, I think your ISY programs are sending too many requests. I see duplicate web requests immediately after each other - it could be pissing the Ecobee servers off. Example: 2018-08-18 19:55:02 - Web Request: /isylink/nodes/n001_upecobee/cmd/GV4/0/25 2018-08-18 19:55:02 - Web Request: /isylink/nodes/n001_upecobee/cmd/GV4/0/25
rccoleman Posted August 20, 2018 Posted August 20, 2018 3 hours ago, io_guy said: I see way more webrequests (from the ISY) than responses from the Ecobee. I'll add some additional logging to the next version to capture each send to the Ecobee (not just response) so they can be matched up to help debug this. Also, I think your ISY programs are sending too many requests. I see duplicate web requests immediately after each other - it could be pissing the Ecobee servers off. Example: 2018-08-18 19:55:02 - Web Request: /isylink/nodes/n001_upecobee/cmd/GV4/0/25 2018-08-18 19:55:02 - Web Request: /isylink/nodes/n001_upecobee/cmd/GV4/0/25 Hmm.. I'm not purposefully sending the same command repeatedly, but I do send some back-to-back commands to adjust both upstairs and downstairs thermostats. I just did a search through all of my programs and I don't see anything weird like that. What does that particular command do? It would help me track it down. Edit: I think I figured out that that command is changing the "fan on time", which I do adjust regularly based on whether there's a meaningful difference in temperature between the various sensors on each floor (basically, I either set it to almost always on - 55 mins/hour - or off - 0 mins/hour). That could potentially change whenever the sensor temperatures update, but I'm reacting more often than necessary (once for every single sensor report that comes in, when it comes in). I just fixed that and will keep an eye on it. Thanks for the tip!
larryllix Posted August 20, 2018 Author Posted August 20, 2018 1 hour ago, rccoleman said: Hmm.. I'm not purposefully sending the same command repeatedly, but I do send some back-to-back commands to adjust both upstairs and downstairs thermostats. I just did a search through all of my programs and I don't see anything weird like that. What does that particular command do? It would help me track it down. Edit: I think I figured out that that command is changing the "fan on time", which I do adjust regularly based on whether there's a meaningful difference in temperature between the various sensors on each floor (basically, I either set it to almost always on - 55 mins/hour - or off - 0 mins/hour). That could potentially change whenever the sensor temperatures update, but I'm reacting more often than necessary (once for every single sensor report that comes in, when it comes in). I just fixed that and will keep an eye on it. Thanks for the tip! Off topic here: Did you know those fan cycle times do not represent anything in real time values? I found 5 minutes runs them about 20% of the time, 10 minutes about 40% of the time, and more than 15 minutes almost on all the time. Somewhere in their development the actual times must have gotten scrambled. I found this out by recording the cycle times on graph charts with ISY's logging (MWareman) capabilities. Still the best fan cycler I have seen in a stat.
rccoleman Posted August 20, 2018 Posted August 20, 2018 10 minutes ago, larryllix said: Off topic here: Did you know those fan cycle times do not represent anything in real time values? I found 5 minutes runs them about 20% of the time, 10 minutes about 40% of the time, and more than 15 minutes almost on all the time. Somewhere in their development the actual times must have gotten scrambled. I found this out by recording the cycle times on graph charts with ISY's logging (MWareman) capabilities. Still the best fan cycler I have seen in a stat. Interesting, no I’ve never paid attention to it. I’d prefer a simple on/off, but this is the best that I can do right now.
larryllix Posted August 20, 2018 Author Posted August 20, 2018 46 minutes ago, rccoleman said: Interesting, no I’ve never paid attention to it. I’d prefer a simple on/off, but this is the best that I can do right now. I run Fan On (100%) in my 3-7 AM "wee hours" climate type for middle of the night stuffiness, and 5 min fan cycles the rest of time I am not on vacation, for economy. ISY switches my main ecobee over to 45 minutes when the temperature outside is extreme only. I love the way you can manipulate preset in ecobee without having to rely on saved temperatures and direct controls. If any ISY control fails, the ecobee stats just continue to do their thing. BTW: I just switched my ecobee4 Alexa off. It has become very annoying, answering the TV and other noises constantly, and hasn't kept up to the other Echo technologies. No messaging, no timer names, no alternate trigger name, it doesn't hear close range hardly at all, competes with other dots for responses, muffled sound, etc. I have enough other Dots anyway.
rccoleman Posted August 20, 2018 Posted August 20, 2018 I already have my house blanketed with Echoes and Dots, so the addition of Alexa to Ecobee4 (and pretty much nothing else) was worthless to me. I'm not sure what more I want in terms of hardware capabilities, but that sure wasn't it.
Recommended Posts
Archived
This topic is now archived and is closed to further replies.