Unavailable after reboot or some time

[Monacoslo]

After HA reboot my devices are Unavailable untill I manually trigger them in eWelink application. Same issue is also after some time not using them even without HA reboot.

Any sollution? Maybe to "ping" them somehow or sth?

[prosymnus]

same here. T1 switch unavailable every time restart the HA.

[LennyLip]

ups, sorry #68

[mantaalex]

im having the issue to hopefully someone could fix it please report someone if fixed :) tx

[tawagny]

same here , I tried to add them again and it worked but again something went wrong

UPDATE :
Ok I think my Router is to blame I just tried with another Device and its working like a charm ,, make sure the multicast is enabled and working properly

[AlexxIT]

All devices unavailable after restart. It does not depend on reload once / always setting.
Devices are automatically detected in the local network after each restart.
If this does not happen, there are some problems with the multicast / router.

[AlexxIT]

@mattsaxon do you know how to force device descovery after HA reboot?
Sometimes devices appear quickly. Sometimes after a few minutes.

[mattsaxon]

I’ve not had this problem with my component. That is likely because it doesn’t use any passed discovery information. I note that your component returns if there is no discovery data passed, but I’m not clear why you do that, since you have all the information from your saved cache or from the ewelink cloud to setup your zeroconf instance. If you did this, it would allow you to setup devices that are currently off. I note that on a restart of HA, devices that are off are not displayed even if they have been previously discovered.

[AlexxIT]

@mattsaxon problem with zeroconf discovery. Devices do not always respond quickly.
Of course, I can save the IP addresses of devices before reboot. But it's not a very nice solution.

[mattsaxon]

I’ve not had any problems with zeroconf response time. I don’t think you should save IPs. I assumed the issue for your component is due to HA not passing the discovery info, not due to zeroconf itself.

EDIT: Just traced through the code and can see where the discovery_info is added, so don't see this as an issue now.

[AlexxIT]

@mattsaxon I think the problem is with the router. For some users, the problem appears more often, for some less often.

mDNS is very complicated traffic.

[mattsaxon]

Agreed. If you implement the approach as discussed In issue #79, I think this will be a lot more acceptable to users as they will see the device greyed out at startup and then come online.

One issue is that any automations however will fail as you can’t automate a device which is unavailable.

A potential approach to fix this is to provide a config setting that states if the device is persistent. ie should never be unavailable. This would allow users to configure if you should always have marked as available in home Assistant and then implement some retry logic to deal with the poor network.

I myself haven’t had to implement the setting in my component, but did implement the retry logic.

[AlexxIT]

Try new version #99

[AlexxIT]

Version 2 should handle all the problems.