Announcement

Collapse
No announcement yet.

HS doesn't update state correctly for missing module

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    HS doesn't update state correctly for missing module

    I send an on with HS to a module that's not even plugged in. HS shows the state is ON. Shouldn't it show UNKNOWN for any module that's not plugged in?

    Also, doing a "Poll Devices" doesn't return UNKNOWN.

    #2
    Right now the command acknoledgement in UPB is a little flakey. Even if the device does not respond the device representation in HS would be On since that is the last command sent out. Same with Z-Wave.

    Not all devices support status reports so if a device does not respond the last state is used.



    Originally posted by DC
    I send an on with HS to a module that's not even plugged in. HS shows the state is ON. Shouldn't it show UNKNOWN for any module that's not plugged in?

    Also, doing a "Poll Devices" doesn't return UNKNOWN.
    Jon Ort
    JonOrt@The--Orts.com
    (Remove the dashes in the address, spam is getting out of hand)

    Comment


      #3
      I thought the purpose of having 2-way devices is so that HS always knows the true state of the device, not just what he thinks it is because of what he last sent.

      Comment


        #4
        Most devices do respond to polling. Most devices do support ACK responses to commands. There are some issues with the PIMs that are currently being worked on with the ACK issues. I suppose an option could be added to set the device status to unknown if at any time a device did not respond as expected, but that is not quite the main purpose of 2-way devices. For instance the lamp modules have local load sense, and they will transmit that the light was turned on locally so HomeSeer is kept in-sync. Also the wall switches transmit when they are controlled locally so again HomeSeer is kept in-sync. Polling should never be needed except at startup to get the initial state of the devices.



        Originally posted by DC
        I thought the purpose of having 2-way devices is so that HS always knows the true state of the device, not just what he thinks it is because of what he last sent.
        Jon Ort
        JonOrt@The--Orts.com
        (Remove the dashes in the address, spam is getting out of hand)

        Comment


          #5
          OK, I understand about the device sending the state when controlled locally, and that does work but not instantly. When I turn a lamp on locally, it takes 30-60 seconds for HS to reflect the state change. I can live with that.

          But, what about when HS sends the device a command. If HS sends an "ON" to a device that doesn't even exist, I would sure hope that HS wouldn't show the device is ON. I'm not saying HS should continuously poll all devices to see if they are on or off but if HS sends ON to a device, and doesn't get any ACK back, he should show UNK as the device state -- because he truly does not know. Unless, of course it's a bug in the ACK processing logic so that the device might actually be on, and sent the ACK command, but HS just didn't see it.

          Would it be possible to make that an option? The choices could be "Set state to last HS command" or "Set actual state received from device".

          Comment


            #6
            30 to 60 seconds is horrible.. I get an update in about 1 second. It should be no where near that long (this is not the HS web UI updating by itself, this is the device itself updating). Once the Ack stuff gets fixed I will add the "Unknown" option.



            Originally posted by DC
            OK, I understand about the device sending the state when controlled locally, and that does work but not instantly. When I turn a lamp on locally, it takes 30-60 seconds for HS to reflect the state change. I can live with that.

            But, what about when HS sends the device a command. If HS sends an "ON" to a device that doesn't even exist, I would sure hope that HS wouldn't show the device is ON. I'm not saying HS should continuously poll all devices to see if they are on or off but if HS sends ON to a device, and doesn't get any ACK back, he should show UNK as the device state -- because he truly does not know. Unless, of course it's a bug in the ACK processing logic so that the device might actually be on, and sent the ACK command, but HS just didn't see it.

            Would it be possible to make that an option? The choices could be "Set state to last HS command" or "Set actual state received from device".
            Jon Ort
            JonOrt@The--Orts.com
            (Remove the dashes in the address, spam is getting out of hand)

            Comment


              #7
              OK, your comment gave me a hint. The device state does get updated very quickly. I was sitting and watching for the screen to update. If I click "status" to refresh the screen immediately after I turn the lamp on, it shows the correct device state as quick as I can click status.

              Sorry about that.

              Comment


                #8
                I loved the immediate feedback of the old HS Windows GUI. HS does update the device "row" when you click to control the device from the web page. I wonder if they can do it whenever a device changes.



                Originally posted by DC
                OK, your comment gave me a hint. The device state does get updated very quickly. I was sitting and watching for the screen to update. If I click "status" to refresh the screen immediately after I turn the lamp on, it shows the correct device state as quick as I can click status.

                Sorry about that.
                Jon Ort
                JonOrt@The--Orts.com
                (Remove the dashes in the address, spam is getting out of hand)

                Comment

                Working...
                X