Still getting could not be assigned a valid UID

[pr4bh4sh]

What is the issue or idea you have?
Issue is described at #738

Does it only happen on a specific device? Please run adb devices -l and paste the corresponding row.
Google Pixel devices(5,6,7,8)

Please provide the steps to reproduce the issue.
Install and uninstall an apk 10000 times.

Describe the solution you'd like
Don't have a solution. However, can we have option of details logging for what caused the install/uninstall action. It's not logged at the moment.

stf/lib/units/device/resources/service.js

Lines 73 to 96 in b5b3f91

	.catch(function() {
	log.info('Installing STFService')
	// Uninstall first to make sure we don't have any certificate
	// issues.
	return adb.uninstall(options.serial, resource.pkg)
	.timeout(15000)
	.then(function() {
	return promiseutil.periodicNotify(
	adb.install(options.serial, resource.apk)
	, 20000
	)
	.timeout(65000)
	})
	.progressed(function() {
	log.warn(
	'STFService installation is taking a long time; ' +
	'perhaps you have to accept 3rd party app installation ' +
	'on the device?'
	)
	})
	.then(function() {
	return getPath()
	})
	})

[jupe]

I’ve been using several pixels succesfully, so I guess there is something wrong in your setup. I could add logs there to help investigation.

But that sounds a bug if stf try to reinstall STFService in a loop..

[pr4bh4sh]

Yeh, it could potentially be bug. I am using latest docker images and not replacing the stf service apk anymore which I did in the linked issue. So ideally the issue shouldn't happen.

Additional logs would be great for investigation and could give me a hint for the root cause.

I have 3 NUC each with 10 pixels, 5,7 and 8. So it's quite annoying issue for me at the moment.

[jupe]

could you describe more about your deployment ? Doesn't you see any warnings/ prints in logcat, what about stf side, e.g. Incompatible version.. etc?

[pr4bh4sh]

I have used https://medium.com/@nikosch86/getting-started-with-automated-in-house-testing-on-android-smartphones-using-stf-dafecee4a8ee to setup the device farm

Here's how I start the provider which is running on the NUC which have Pixel physically connected

sudo docker run --name docker_devicefarmer_android_provider -d --net=host --restart unless-stopped -p 7400-7700:7400-7700 -e "ZMQ_TCP_KEEPALIVE=1" -e "ZMQ_TCP_KEEPALIVE_IDLE=300" devicefarmer/stf:latest stf provider --name devicefarmhub2 --connect-sub tcp://<devicefarm_ui_host>:7250 --connect-push tcp://<devicefarm_ui_host>:7270 --storage-url http://<devicefarm_ui_host>:8082/ --heartbeat-interval 10000 --screen-ws-url-pattern "ws://<devicefarm_ui_host>:8082/d/devicefarmhub2/<%= publicPort %>/" --adb-port 9128 --min-port 7400 --max-port 7700 --screen-jpeg-quality 10

Now that you said about warning and logs, I don't know if i'm logging them correctly.

[pr4bh4sh]

So, I finally figured out how to see the stf logs, will monitor overtime and report back if I notice something. I have already formatted the phones so it would take probably around ~3 month for issue to appear.

for anyone reading this(and future self), command for checking logs is
docker logs [stf-container-id]