I have GGR connected to a physical server running Appium and two devices connected to it. Appium is capable of running a test each on the two devices, upon a request routed by GGR. Should I be making two entries in GGR, one for each device name?

In the code for the test, we don't want to give the browser name capability.

Thanks. This has no mention of device or platform?

there were no device and platform when document was created, all browsers ran on linux, there were no support for mobile devices.

Oh ok. But now these capabilities are avbl right?

via capabilities, ggr just proxying to upstream host

Yes, however I would like to proxy all Android requests to a single Appium server. Instead of specifying and actual device name each time

I would let the Appium server distribute them to various devices

здравствуйте, не подскажите, локально поднимаю селеноид все хорошо, удаленно на машине развернул девопс. как я понял ему нужно еще image browsers прикрутить

copy paste respective host in XML

docker pull пусть сделает всем образам из browsers.json

и полетит

Попытаемся

Hi Guys, I love the selenoid project, everything is working great for me for Chrome and Firefox. I'm trying to add also edge in my linux mint, but I'm stuck and I need your expertise!

On selenoid ui I can see the edge browser that I added.
I'm using the port 4444 for edge,inside the file browsers.json which is in /root/aerokube/selenoid/browsers.json

I installed windows on qemu, I can login via VNC.
I disabled the windows firewall and all three layers under Windows Security.

On file entrypoint.sh I configured ports like this:

sudo qemu-system-x86_64 -enable-kvm \
        -machine q35 -smp sockets=1,cores=2,threads=4 -m 4096 \
        -usb -device usb-kbd -device usb-tablet -rtc base=localtime \
        -net nic,model=virtio -net user,hostfwd=tcp::4442-:4444 \
        -drive file=snapshot.img,media=disk,if=virtio \
        -loadvm windows

To start selenoid, I'm doing sudo ./cm start selenoid

To run the image: sudo docker run -it --rm --privileged -p 4442:4444 -p 5900:5900 windows/edge:18
On windows CMD I'm running:
C:\Windows\System32\IEDriverServer.exe --host=0.0.0.0 --port=4444 --log-level=DEBUG

When I run curl like this:

curl -X POST http://127.0.0.1:4444/session -d '{"capabilities": {"browserName": "MicrosoftEdge", "platform": "win10"}}'

I have an error: curl: (52) Empty reply from server


I'm running curl from my linux box, not from windows inside qemu, right?

everything looks correct, try to find something in edge logs via vnc

Пытаюсь разгадать загадку. Настроил ggr, распределение по серверам в квоте одно, а реальная картина использования браузеров совсем другая. Может кто знает, куда можно копать? В логах ggr не вижу никаких ошибок. На графике как раз кол-во "used" браузеров из / status

здравствуйте. вопросы по запуску тестов для нативного приложения андроида. как я поняла, на маке вариант только ставить вмваре с линуксом и запускать в нем селеноид. делала так https://www.youtube.com/watch?v=RmR2Z7siiNM&t=63s. контейнер запустился, но без эмулятора. по PS ax эмулятора нет. kvm-ok - ok. есть может какие-то тонкости при создании вм, которые я не учла?

I made a step further, now If I manually launch qemu with this command:

sudo qemu-system-x86_64 -enable-kvm \
        -machine q35 -smp sockets=1,cores=2,threads=4 -m 4096 \
        -usb -device usb-kbd -device usb-tablet -rtc base=localtime \
        -net nic,model=virtio -net user,hostfwd=tcp::4442-:4444 \
        -drive file=snapshot.img,media=disk,if=virtio \
        -loadvm windows

I can launch edge with a CURL.

If instead I build a docker image, with the same above command inside entrypoint.sh, the CURL fails

This is probably due to the fact that I edited DockerFile. In order to match the same qemu version installed on my machine, I put Ubuntu 20.04 in the DockerFIle and instead of installing qemu I installed qemu-system. This package installs also the mandatory qemu-system-x86_64 package

#DockerFile:
FROM ubuntu:20.04
RUN apt update && \
    apt -y upgrade && \
    apt -y install xvfb x11-utils x11vnc qemu-system
COPY hdd.img /
COPY snapshot.img /
COPY entrypoint.sh /
ENTRYPOINT ["/entrypoint.sh"]


Any suggestion on how to proceed? Thanks