Jump to content

Welcome to Smart Home Forum by FIBARO

Dear Guest,

 

as you can notice parts of Smart Home Forum by FIBARO is not available for you. You have to register in order to view all content and post in our community. Don't worry! Registration is a simple free process that requires minimal information for you to sign up. Become a part of of Smart Home Forum by FIBARO by creating an account.

 

As a member you can:

  •     Start new topics and reply to others
  •     Follow topics and users to get email updates
  •     Get your own profile page and make new friends
  •     Send personal messages
  •     ... and learn a lot about our system!

 

Regards,

Smart Home Forum by FIBARO Team


  • 0

Fibaro HC2 Crashed??


morpheus75

Question

I dont know if anyone else has come accross this, but this morning my system seemed to have froze! My last scene that ran was about 00:05, and after that i have just a couple of scenes to turn off bedroom lamps and these didnt work, neither did the morning water heating, central heating scenes etc

when i tried to manually run them nothing worked. tried to disarm sensors but they would auto-re arm. tried the dimmers and they would stay on. no scene would run.

a restart of HC2 did the trick.

a client of mine reported the same thing this morning so Question: Has anyone else experienced this, this morning?

 

thanks

Link to comment
Share on other sites

Recommended Posts

  • 0

Since I have System status VD I can monitor what is happening with my HC2. For last five weeks my HC2 was working without any glitches. What I have on my system you can see form my signature. I believe that I can freely say it is big system with 96 included physical devices and numerous scenes and VD's running. I noticed that after three consecutive weeks running memory usage stabilized at 26% free (used at 44%, buffered at 6% and cached at 24%) while processors usage jumps between 10% and 80% depending on number of processes running.

 

Since I was connecting remotely during this last five weeks there was only problem with web GUI refreshing and sometimes I also get my home page empty like on your picture @morpheus75. Simple click on the refresh button on the browser in most cases helped but sometimes I had to reconnect.

 

I rebooted my HC2 yesterday after all this time working without glitches because I included some more devices to z-wave network and that caused memory usage to jump to the value when I get warning (20% free), but still not asking for reboot (15% free) or forced reboot (10% free). I always do that as preventive step and not as necessary one.  My pooling devices settings is at default so for most of them disabled. Also I have disabled marking non battery powered devices as dead. All my code is always first checking last reported status of the devices before sending command to them so not to cause any unnecessary z-wave traffic.

 

I'm still trying to figure out why my HC2 get frozen on 4.120. It happened when my wife activated scene for disarming alarm and turn on lights. This scenes never finished their job so she came back home in dark and armed sensors but never get alarm activated due to frozen HC2. I checked my code and run it many times but never get HC2 frozen again. BTW arming and disarming scenes handles many sensors and sends many commands over the network so I suspect that most probably z-wave chip get stuck but I have no tools or any means to prove that theory.

 

And just to mention, from my experience if you use api.get, api.put and api.post functions in sequence then put delay of at least 500ms between them. I have one code that was failing on this functions until I used this delay between them. Also from my experience if you send command to a device and then you want to check it's state change then you need to have at least delay of 100 to 200ms between otherwise you will get wrong state. Also when you are using TCP ports to communicate to other devices on the network don't forget to close port after job is done otherwise you can end up with device not responding any more to HC commands and in some cases even 503 errors.

 

Sorry for long post, there are more things that I feel needs to be pointed out, but maybe next time. HC system doesn't have any means of self protection from bad programming or wrong procedures so one cook book would be nice to have. Fibaro team?

  • Thanks 1
Link to comment
Share on other sites

  • 0
  • Inquirer
  • 2 minutes ago, Lambik said:

    Just a remark, check if there any devices with (<5%) low battery or ones who are reporting 'dead'. In the past I noticed if a device is dead and is used by the zwave network as a hop to other devices or when a dying device is connected to a (stand-alone) hop, the HC2 freezes (because the hop has problems connecting or overloaded with connecting data).

     

    2 weeks ago I got a strange message from a device reporting low battery. I didn't had the time to check so I disabled the mail messages. After a day I checked and the device wasn't reporting low battery any more. After some investigation I found that a nearby device was gone 'dead' (because of low battery) and probably polluted z-wave communications somehow.

     

    I cannot prove the above assumptions but it is something I noticed. The HC2 can be get alive again after some period of time, I assume because of an (automatic initiated) network 'heal'.

     

    Maybe you can check batteries just to rule out my assumption. Just to be sure that I'm wrong. ;)

     

    thanks for your input.

    from memory battery operated devices dont act as repeaters.

    also all my battery operated sensors have all got good charge so unfortunately i have ruled this out.

     

    @AutoFrank has given me a VD, which i am currently using and am monitoring the core cpu usage.

    when i get a chance will amend the code and get it to plot a graph on thingspeak and make public so everyone can see. hopefully should be able to work out firstly when it actually crashed and secondly will have a look at scenes to see what was active at that time.

    thanks again for your thought

     

    Link to comment
    Share on other sites

    • 0
    25 minutes ago, morpheus75 said:

    thanks for your input.

    from memory battery operated devices dont act as repeaters.

    also all my battery operated sensors have all got good charge so unfortunately i have ruled this out.

     

    I'm aware that battery operated devices don't act as repeaters (hop). I tried to point out that repeaters could act strange when a battery device connected to the repeater dies.

    But if all your devices have good charge then this can be ruled out. (Unless the network has another kind of interference)

     

    Happy hunting! ;D

     

    EDIT/Merged:

    Just curious, when the HC freezes:

    1 - Which LED's are lit? (Is zwave network communicating, is TCP/IP network connected)

    2 - Are you able to log in into the webinterface? (Did the webserver crashed)

    3 - Are you able to ping the HC2?

    Another questions:

    4 - Is your network settings set to DHCP or manually?

    5 - Do you have root access?

     

    Another thing you could check: Time settings (is your actual time/date drifted)

    Edited by Lambik
    Link to comment
    Share on other sites

    • 0
  • Inquirer
  • 44 minutes ago, Lambik said:

     

    I'm aware that battery operated devices don't act as repeaters (hop). I tried to point out that repeaters could act strange when a battery device connected to the repeater dies.

    But if all your devices have good charge then this can be ruled out. (Unless the network has another kind of interference)

     

    Happy hunting! ;D

     

    EDIT/Merged:

    Just curious, when the HC freezes:

    1 - Which LED's are lit? (Is zwave network communicating, is TCP/IP network connected)

    2 - Are you able to log in into the webinterface? (Did the webserver crashed)

    3 - Are you able to ping the HC2?

    Another questions:

    4 - Is your network settings set to DHCP or manually?

    5 - Do you have root access?

     

    Another thing you could check: Time settings (is your actual time/date drifted)

     

    i understand now re battery operated devices. thanks

     

    re questions:

    1. normal 3 LED lit

    2. web cannot login see pic in first post - i even clear cache etc Didnt try phone app, will try next time

    3. not tried to ping will try next time

    4. fixed static IP given - most of my network has fixed IP's

    5. root access? no

     

    On another note i have added the Thingspeak graphs so link here to monitor the Core 1 & Core 2 memory usage:

     

    Please login or register to see this link.

     

    Also there was a discussion on 'Polling' so i have gone through all my devices and disabled polling on all devices that i didnt need. so lets see what happens.

    cheers

    Link to comment
    Share on other sites

    • 0

    1 - Did you check your time setting?

    2 - Are you able to set the network setting to DHCP and link the HC2 mac-address to a reserved IP?

    If so, try to configure also all devices which the HC2 communicate with by DHCP and reserved mac-addressing.

    3 - When your HC freezes again, check the zwave LED for communications (by activating a motion/door/window sensor for example).

    4 - About polling, if you configure global polling settings right, there is no need to overriding polling settings by device.

    5 - When HC freezes, use another webbrowser to check if you experience the same behaviour. Even better, check using an other computer.

     

    I will check your monitoring graphs once in a while to see if some weird values shows up.

    Edited by Lambik
    Link to comment
    Share on other sites

    • 0
    2 hours ago, Lambik said:

     

    2 weeks ago I got a strange message from a device reporting low battery. I didn't had the time to check so I disabled the mail messages. After a day I checked and the device wasn't reporting low battery any more. After some investigation I found that a nearby device was gone 'dead' (because of low battery) and probably polluted z-wave communications somehow.

     

     

     

    9 hours ago, Jamie mccrostie said:

     

    Same here no new scenes or devices for over a month.... I did have a fibaro sensor go off line two days later flat battery no battery warning just stopped working 

    had to exclude and reinclude to get it working

    All seems to ok since

    In my case maybe, but why no battery alarm?

    2 hours ago, Sankotronic said:

    one cook book would be nice to have. Fibaro team?

     

    Yep

    Link to comment
    Share on other sites

    • 0
    14 hours ago, Lambik said:

    Just a remark, check if there any devices with (<5%) low battery or ones who are reporting 'dead'. In the past I noticed if a device is dead and is used by the zwave network as a hop to other devices or when a dying device is connected to a (stand-alone) hop, the HC2 freezes (because the hop has problems connecting or overloaded with connecting data).

     

    2 weeks ago I got a strange message from a device reporting low battery. I didn't had the time to check so I disabled the mail messages. After a day I checked and the device wasn't reporting low battery any more. After some investigation I found that a nearby device was gone 'dead' (because of low battery) and probably polluted z-wave communications somehow.

     

    I cannot prove the above assumptions but it is something I noticed. The HC2 can be get alive again after some period of time, I assume because of an (automatic initiated) network 'heal'.

     

    Maybe you can check batteries just to rule out my assumption. Just to be sure that I'm wrong. ;)

     

    I have been thinking that too. I'm pretty certain the old, round "Aeotec Multisensor" can cause troubles, when batteries are low. It probably starts sending data at a certain point until batteries are depleted. I have actually seen sensor with the red blinking light, that would no longer shut down. Using rechargeable NiMH cells improves this situation, probably because this battery chemistry goes from a constant voltage (1.2 V) to dead (1.0 V) very, very quickly. So quick in fact, that the sensor often won't be able to send a "battery empty" report. Also, when cells have unequal capacities or you leave them in too long (about a week I think) theses sensors will deteriorate your cells, because they keep discharging the batteries below 1.0 V (per cell). You can somewhat revive dead cells by cycling them, but that takes a lot of time (can take up to 5 charge/discharge cycles).

    Edited by petergebruers
    Link to comment
    Share on other sites

    • 0
    On 23-7-2017 at 8:34 AM, petergebruers said:

     

    I have been thinking that too. I'm pretty certain the old, round "Aeotec Multisensor" can cause troubles, when batteries are low. It probably starts sending data at a certain point until batteries are depleted. I have actually seen sensor with the red blinking light, that would no longer shut down. Using rechargeable NiMH cells improves this situation, probably because this battery chemistry goes from a constant voltage (1.2 V) to dead (1.0 V) very, very quickly. So quick in fact, that the sensor often won't be able to send a "battery empty" report. Also, when cells have unequal capacities or you leave them in too long (about a week I think) theses sensors will deteriorate your cells, because they keep discharging the batteries below 1.0 V (per cell). You can somewhat revive dead cells by cycling them, but that takes a lot of time (can take up to 5 charge/discharge cycles).

     

    FYI

     

    I have several of these devices also. In the past I used rechargeable batteries also, but switched back to normal batteries because of the short time I had to change/recharge them. I found that the default HC2 device settings are the best settings to use and using non-battery powered devices in the neighbourhood (acting as extender) did the trick.

     

    When the led keeps blinking it means (as far as I noticed) the connection to your zwave-network has been lost. A triple press on the inside button reconnects it. I have also seen this behaviour on non-battery powered Aeotec Multisensors.

     

    The round Aeotec Multisensors (DSB05-ZWEO) are available in several firmware versions. I had a lot of problems with the oldest firmware versions, the 'newer' seems to be better.

     

    An other thing is that, when you have an area where there are zwave connecting problems, Fibaro's FGK-101/107 battery-powered door/window sensors drain very fast.

     

    For monitoring these kind of behaviours I make use of the

    Please login or register to see this link.

    , installed on a LAN-only webserver (for security reasons), so I can check what is happening.

     

    Link to comment
    Share on other sites

    • 0
    2 minutes ago, Lambik said:

    I have several of these devices also. In the past I used rechargeable batteries also, but switched back to normal batteries because of the short time I had to change/recharge them. I found that the default HC2 device settings are the best settings to use and using non-battery powered devices in the neighbourhood (acting as extender) did the trick.

    Confirmed. If I take care of my NiMH cells, I can get 4 -8 weeks in the old round MultiSensor. I can't get decent battery life from a 400 mAh LiFePO4 cell in a MS6 though. I gave up on diagnosing this, so there might be a solution, but I get from other posters MS6 really shines as USB powered device. I did a lot of tests regarding reporting interval, could get what I wanted.

     

    4 minutes ago, Lambik said:

    When the led keeps blinking it means (as far as I noticed) the connection to your zwave-network has been lost. A triple press on the inside button reconnects it. I have also seen this behaviour on non-battery powered Aeotec Multisensors.

    Confirmed, with a twist. I had one of the round sensors on my desk (line of sight to HC 6 meter). I know its batteries where poor. Then it started blinking. It continued to do so until the battery was dead. So when it does that, it kills your rechargeable cells.

     

    7 minutes ago, Lambik said:

    An other thing is that, when you have an area where there are zwave connecting problems, Fibaro's FGK-101/107 battery-powered door/window sensors drain very fast.

    Confirmed. Speculation: poor connection drains cel faster than normal. Low or poor battery causes communication to fail and it get's stuck in a loop. I've seen some of them do what the MultiSensor does: blink the LED until the battery dies.

    Link to comment
    Share on other sites

    • 0

    @petergebruers Thanks for the feedback. Nice to see we noticed some same behaviours and it's not just motivated by those goofy voices in my head... ;D

    Link to comment
    Share on other sites

    • 0
  • Inquirer
  • hello all

    its been 3 days now and the system is functioning properly again. so far no crashes etc

    the CPU monitoring is going well and it looks like it stays around 5-10% and occasionally jumps to 30-40% for a few seconds before stabilising. Thanks to @AutoFrank

    the changes i have made are:

     

    update to 4.134 (thought i would try.......)

    polling has been disabled on the non-essential relays

     

    thats it. no other changes. battery level on my Fibaro Motion Sensors are all still good and the keyfob. I dont have any other battery devices. The SP103 and Aeotec 4 in 1 are all powered using USB and/or power supplies

    and are working fine.  So i am stumped as to what may have caused. im going to leave CPU monitoring on and data to thingspeak on.

     

    thanks all

     

    Link to comment
    Share on other sites

    • 0

    Good to hear @morpheus75!

     

    Just one question though, did you change default polling settings on devices? Why, if I may ask?

    Link to comment
    Share on other sites

    • 0

    @morpheus75

    Great to hear its stabilized but not good that root cause is still unknown

     

    BTW - I went and disabled all the polling on the majority of my devices and the system went all ga ga, lots of dead devices

    I had to recover from a previous backup. I don't know the reason why but would caution against a gross change in polling settings across a system

    Link to comment
    Share on other sites

    • 0
  • Inquirer
  • 18 minutes ago, Lambik said:

    Good to hear @morpheus75!

     

    Just one question though, did you change default polling settings on devices? Why, if I may ask?

     

    there is a thread on here regarding polling. all new devices that are added to HC2 are defaulted 'device excluded from polling' so the consensus was to reduce z-wave traffic.

    so i excluded all my dimmers and relays from polling and just left the ones that were important for me to default polling. all battery devices dont have that option so i left them alone.

    the aeotec sensor remained at 2 mins sending data and HEM both set to send data every 2 mins aswell.

    so far no dead devices and no system crashes etc etc so im assuming that has reduced z-wave traffic and therefore potentially prevented any crashes...

     

     

    7 minutes ago, AutoFrank said:

    @morpheus75

    Great to hear its stabilized but not good that root cause is still unknown

     

    BTW - I went and disabled all the polling on the majority of my devices and the system went all ga ga, lots of dead devices

    I had to recover from a previous backup. I don't know the reason why but would caution against a gross change in polling settings across a system

     

    mine has been OK. all dimmers, relays have been excluded from polling.  only things i have left on are:

    tkb plugs . x 3

    swiid switch . x 3

    UBS x 5

    aeotec 4 in 1 x 1

    HEM x 2

    rest all disabled.

    so far so good...

     

    Link to comment
    Share on other sites

    • 0
    6 minutes ago, morpheus75 said:

     

    there is a thread on here regarding polling. all new devices that are added to HC2 are defaulted 'device excluded from polling' so the consensus was to reduce z-wave traffic.

    so i excluded all my dimmers and relays from polling and just left the ones that were important for me to default polling. all battery devices dont have that option so i left them alone.

    the aeotec sensor remained at 2 mins sending data and HEM both set to send data every 2 mins aswell.

    so far no dead devices and no system crashes etc etc so im assuming that has reduced z-wave traffic and therefore potentially prevented any crashes...

    Hmmm, completely missed that topic....o.O:oops:

     

    As far as my experience goes, I never had polling problems using the default recommended global settings (Z-Wave Network -panel) and 'Devices uses global polling queue'. I will try to find/read the topic about that.....

    Link to comment
    Share on other sites

    • 0
    18 minutes ago, morpheus75 said:

     

    there is a thread on here regarding polling. all new devices that are added to HC2 are defaulted 'device excluded from polling' so the consensus was to reduce z-wave traffic.

    so i excluded all my dimmers and relays from polling and just left the ones that were important for me to default polling. all battery devices dont have that option so i left them alone.

    the aeotec sensor remained at 2 mins sending data and HEM both set to send data every 2 mins aswell.

    so far no dead devices and no system crashes etc etc so im assuming that has reduced z-wave traffic and therefore potentially prevented any crashes...

     

     

     

    mine has been OK. all dimmers, relays have been excluded from polling.  only things i have left on are:

    tkb plugs . x 3

    swiid switch . x 3

    UBS x 5

    aeotec 4 in 1 x 1

    HEM x 2

    rest all disabled.

    so far so good...

     

     

    That's good news @morpheus75, I might try again soon. I moved my Hc2 (about 6 foot into a different room) and added a bigger antennae and letting the system settle for a while

     

    Link to comment
    Share on other sites

    • 0
    On 7/21/2017 at 8:55 PM, morpheus75 said:

    Please login or register to see this attachment.

     

    Ive had the exact same error, HC2 version is V 4.130.

     

    Could not control HC2 with computer via browser (safari) (page was same as yours) 

    Could not control HC2 with Fibaro app on IOS, tried different devices.

    Scenes with motion sensors stopped working. HC2 was completley "stuck"

    3 LEDs lit (power, ethernet, internet),

    Z-wave indication LED blinking "normally" indication communication with devices.

    Static IP set.

    No root acess.

     

    This happened during the night while sleeping so very little going on with the HC2 except polling and temp/lux reports.

     

    Only solution was to power down with power button.

    After restart the HC2 got stuck in booting all LEDs where ON and the wrench LED was constantly blinking.

    Did a second restart with the power button and the HC2 started normally after that.

     

    I have not had this problem before, this was my first time using 4.130, this did not happen during V4.120

    Global polling enabled with setting 500s, I only have a "small-ish" system with 42 devices and all devices included in the polling queue.

    Now up and running again for 3 days without problems 46% free RAM and very little CPU usage < 10% normally .

     

    @Lambik 

    Some more info.

    Here is the power report from when the HC2 crashed, got stuck, maybe this brings more info.

    Even though the z-wave light was flickering nothing was being recorded from devices.

    Please login or register to see this attachment.

     

    Edited by speedy
    Link to comment
    Share on other sites

    Join the conversation

    You can post now and register later. If you have an account, sign in now to post with your account.

    Guest
    Answer this question...

    ×   Pasted as rich text.   Paste as plain text instead

      Only 75 emoji are allowed.

    ×   Your link has been automatically embedded.   Display as a link instead

    ×   Your previous content has been restored.   Clear editor

    ×   You cannot paste images directly. Upload or insert images from URL.

    ×
    ×
    • Create New...