Hi Lava users,
I am using the sample you sent me last week to run tradefed: https://gist.github.com/mwasilew/e6e80a351d43388b1c3342098012ec
Until we have a real device to run on Tradefed we decided to use it with qemu, so the question is:
1. How do we deploy android on qemu via Lava?
2. How to boot android via lava?
Do you have any sample job for the above questions?
george
________________________________
From: George Nistor
Sent: Thursday, August 8, 2019 2:19:05 PM
To: Vladut Magas <vladut.m(a)l4b-software.com>; Inon Sharony <inon.s(a)l4b-software.com>; Yair Podemasky <yair.p(a)l4b-software.com>
Cc: Andreas Schl?ter <andreas.s(a)l4b-software.com>; Ilya Fedusiv <ilya.f(a)l4b-software.com>
Subject: FW: [Lava-users] running commands locally from a Lava job
Here is the answer I have just received from Linaro via their mail-list.
They say it is not a good idea to build tradefed with lava, but it can be used in the scenario to run TF from an LXC container - they have said.
Tradefed should be built from outside Lava.
george
-----Original Message-----
From: Milosz Wasilewski <milosz.wasilewski(a)linaro.org>
Sent: Thursday, August 8, 2019 2:07 PM
To: George Nistor <george.n(a)l4b-software.com>
Cc: lava-users(a)lists.lavasoftware.org
Subject: Re: [Lava-users] running commands locally from a Lava job
On Thu, 8 Aug 2019 at 11:50, George Nistor <george.n(a)l4b-software.com> wrote:
>
> Thanks for the answers,
>
> 1. What do you mean by "with LXC container acting as a host side ". It is an LXC container on the machine where Lava runs and inside this containes runs Tradefed?
yes, exactly. Here is an example job definition:
https://gist.github.com/mwasilew/e6e80a351d43388b1c3342098012ec5c
I can't share the actual job and logs as there are passwords included.
milosz
> 2. Where can I find an example of these Jobs? Could you help me with an link, or sample jobs?
>
> george
>
> -----Original Message-----
> From: Milosz Wasilewski <milosz.wasilewski(a)linaro.org>
> Sent: Thursday, August 8, 2019 1:29 PM
> To: George Nistor <george.n(a)l4b-software.com>
> Cc: lava-users(a)lists.lavasoftware.org
> Subject: Re: [Lava-users] running commands locally from a Lava job
>
> I would say it's a bad idea. LAVA is not meant for building stuff. It can do that but you will need to hack around quite a lot. Secondly IIUC you need to have dependency between build and run steps. If you want to implement this using LAVA this implies multi-node job. There is no other way LAVA will manage dependencies and timings between executions. So this is getting pretty complicated already.
>
> LAVA is already running tradefed jobs (CTS and VTS) using real hardware and it's a pretty common use case. There are 2 implementations of this. One is using single node job with LXC container acting as a host side. Other is using multinode and can shard tests between multiple devices. However tradefed packages are built outside of lava and downloaded before job starts.
>
> To sum up, I think what you're trying to do is not best suited for LAVA. However it might be possible given some effort spent on experimenting.
>
> milosz
>
> On Thu, 8 Aug 2019 at 11:20, George Nistor <george.n(a)l4b-software.com> wrote:
> >
> > Yes
> > For one job to build tradefederation from inside a docker container.
> > For the second job to run Google TradeFederation from the docker container.
> >
> > Would be that possible with Lava?
> >
> > george
> >
> > -----Original Message-----
> > From: Milosz Wasilewski <milosz.wasilewski(a)linaro.org>
> > Sent: Thursday, August 8, 2019 1:16 PM
> > To: George Nistor <george.n(a)l4b-software.com>
> > Cc: lava-users(a)lists.lavasoftware.org
> > Subject: Re: [Lava-users] running commands locally from a Lava job
> >
> > George,
> >
> > Do I understand correctly you would like to use LAVA as a building tool?
> >
> > milosz
> >
> > On Thu, 8 Aug 2019 at 10:47, George Nistor <george.n(a)l4b-software.com> wrote:
> > >
> > > Hi Lava users,
> > >
> > >
> > >
> > > Is it possible to run commands inside a lava job test-section, locally, from a Lava Job?
> > >
> > > Ex:
> > >
> > > cd /media/sh_shared_folders
> > >
> > > docker build -t tradefed .
> > >
> > >
> > >
> > > We have the following setup:
> > >
> > > A machine with Linux Ubuntu;
> > > an Oracle Virtual Box VM on the machine with a Xubuntu distro
> > > image inside the VM we have created 3 Docker containers:
> > >
> > > one with the Jenkins
> > > second one with Lava
> > > third is created for the purpose of running Google TradeFederation
> > > against the Android device
> > >
> > >
> > >
> > > (The AOSP source tree is available to the VM via shared_folders
> > > and the docker container has also access to it. (via docker run -v
> > > dir_host:dir_container) )
> > >
> > >
> > >
> > > What we what to achieve is to have 2 jobs in Lava :
> > >
> > > one for building the Docker image for the purpose of building
> > > TradeFederation inside
> > >
> > > Ex. Running these commande locally:
> > >
> > > source
> > > ./build/make/envsetup.sh
> > >
> > > lunch
> > > <device-target>
> > >
> > > make
> > > tradefed-all -j8
> > >
> > >
> > >
> > > second using the above image, instantiate, create a container
> > > which should run TradeFederation against the android device
> > > connected to the machine
> > >
> > >
> > >
> > > What I need is to be able to run commands locally inside the VM (on Xubuntu distro) from a Lava job. For these we have to somehow not send the command to qemu but run them locally.
> > >
> > > Is this possible? If yes, how can we achieve this?
> > >
> > >
> > >
> > > George Nistor
> > >
> > > _______________________________________________
> > > Lava-users mailing list
> > > Lava-users(a)lists.lavasoftware.org
> > > https://lists.lavasoftware.org/mailman/listinfo/lava-users
> _______________________________________________
> Lava-users mailing list
> Lava-users(a)lists.lavasoftware.org
> https://lists.lavasoftware.org/mailman/listinfo/lava-users
Hi Lava users,
Is it possible to run commands inside a lava job test-section, locally, from a Lava Job?
Ex:
cd /media/sh_shared_folders
docker build -t tradefed .
We have the following setup:
* A machine with Linux Ubuntu;
* an Oracle Virtual Box VM on the machine with a Xubuntu distro image
* inside the VM we have created 3 Docker containers:
* one with the Jenkins
* second one with Lava
* third is created for the purpose of running Google TradeFederation against the Android device
(The AOSP source tree is available to the VM via shared_folders and the docker container has also access to it. (via docker run -v dir_host:dir_container) )
What we what to achieve is to have 2 jobs in Lava :
* one for building the Docker image for the purpose of building TradeFederation inside
Ex. Running these commande locally:
source ./build/make/envsetup.sh
lunch <device-target>
make tradefed-all -j8
* second using the above image, instantiate, create a container which should run TradeFederation against the android device connected to the machine
What I need is to be able to run commands locally inside the VM (on Xubuntu distro) from a Lava job. For these we have to somehow not send the command to qemu but run them locally.
Is this possible? If yes, how can we achieve this?
George Nistor
Hello,
I'm asking if the corresponding lava-slave is still working properly. You
can see that on the web interface in the workers page.
By the way, which version are you using?
Le mar. 6 août 2019 à 16:25, Gumansingh, Smita <Smita_Gumansingh(a)mentor.com>
a écrit :
> Yes Remi,
>
> Corresponding testjob is still running on device from last 3 days after it
> got canceled it from GUI. Testjob state is showing canceling
> and my device state is showing Running.
>
> test-hardware | x86 | running | 15696
>
>
> Thanks & Regards,
> Smita Gumansingh
> ------------------------------
> *From:* Remi Duraffort <remi.duraffort(a)linaro.org>
> *Sent:* Tuesday, August 6, 2019 6:26 PM
> *To:* Gumansingh, Smita
> *Subject:* Re: [Lava-users] Testjob state is cancelling
>
> Hello,
>
> are you sure that the corresponding lava-slave is still running?
>
>
> Rgds
>
> Le mar. 6 août 2019 à 09:11, Gumansingh, Smita <
> Smita_Gumansingh(a)mentor.com> a écrit :
>
>> Hi all,
>>
>> I cancelled a test job but it still running on my device and showing
>> testjob state as “cancelling”. Is there any way to restart the device ,so
>> that I can submit some other job .
>>
>> Killing the specific job id on lava-master using lava-cli also not
>> solving the problem.
>>
>>
>>
>> *Regards,*
>>
>> *Smita Gumansingh*
>>
>>
>>
>> [image: cid:image001.jpg@01D2E5A8.A6033460]
>>
>>
>>
>> Senior test Engineer, EPS PBU Software QA/Dev Ops
>>
>> Mentor Graphics,
>>
>> Nalapad Brigade Centre, Unit No:301 & 302
>>
>> Mahadevapura, Bengaluru, Karnataka 560048, India
>>
>> Office: +91 8067624338 |Cell: +91 9538349245
>>
>> www.mentor.com
>>
>>
>> _______________________________________________
>> Lava-users mailing list
>> Lava-users(a)lists.lavasoftware.org
>> https://lists.lavasoftware.org/mailman/listinfo/lava-users
>
>
>
> --
> Rémi Duraffort
> LAVA Team, Linaro
>
--
Rémi Duraffort
LAVA Team, Linaro
Hi all,
I'm running CTS tests on Android 9, with LAVA.
I noticed the CTS module CtsUsbTests failed because of the error:
"usb serial != adb serial expected:<[]> but was:<[a266fa0a5c85d6ee]>"
This module is a hostsidetest, and checks if usb serial matches adb serial
by running "lsusb -v" on the host and checking the iSerial returned by the
command. (
https://android.googlesource.com/platform/cts/+/refs/heads/master/hostsidet…
)
And this test fails when I run it with LAVA so I checked on my local PC and
the test passes. Indeed, if I run the command "lsusb -v" I get this output
for iSerial:
iSerial 3 351d5a0a5c85d936
The third field is the adb serial, so it matches well.
So I suspected the LXC container to not give the usb serial. So I checked
directly in the container and after running "lsusb -v" I got:
iSerial 3
The adb serial is missing.
So, do you know a way to make the LXC see correctly the usb serial ?
Regards,
Axel
Hello,
We've been using uboot-ums deploy mechanism for quite some time but we really want to move away from dd and star using bmap-tools.
I guess lava does not support it (yet) but I was wondering if there is a way where we can use uboot-ums with bmap-tools.
Thanks
--
Diego Russo | Staff Software Engineer | Mbed Linux OS
ARM Ltd. CPC1, Capital Park, Cambridge Road, Fulbourn, CB21 5XE, United Kingdom
http://www.diegor.co.uk - https://os.mbed.com/linux-os/
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
Right, it is not the first time I ask a similar question (I did ask it last February) and a lava issue has been created: https://git.lavasoftware.org/lava/lava/issues/234
Unfortunately there are no progress on this and actually it has been removed from the milestones.
Can we ask to reconsider it? (
Cheers
-----Original Message-----
From: Lava-users <lava-users-bounces(a)lists.lavasoftware.org> on behalf of Diego Russo <Diego.Russo(a)arm.com>
Date: Tuesday, 6 August 2019 at 13:32
To: "lava-users(a)lists.lavasoftware.org" <lava-users(a)lists.lavasoftware.org>
Subject: [Lava-users] uboot-ums with bmap-tools
Hello,
We've been using uboot-ums deploy mechanism for quite some time but we really want to move away from dd and star using bmap-tools.
I guess lava does not support it (yet) but I was wondering if there is a way where we can use uboot-ums with bmap-tools.
Thanks
--
Diego Russo | Staff Software Engineer | Mbed Linux OS
ARM Ltd. CPC1, Capital Park, Cambridge Road, Fulbourn, CB21 5XE, United Kingdom
http://www.diegor.co.uk - https://os.mbed.com/linux-os/
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
_______________________________________________
Lava-users mailing list
Lava-users(a)lists.lavasoftware.org
https://lists.lavasoftware.org/mailman/listinfo/lava-users
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
Hello LAVA experts,
I have a device which can boot up to android/linux.
The device is connected with two USB cables.
One USB cable is used for flashing images (fastboot flash) and doing some adb shell commands. Via another USB cable, I can use minicom to check uart logs and directly interact with the device (/dev/ttyUSB0).
Can you share some examples (jobs, job definitions) or configurations (like ser2net conf) or any suggestion for doing below test steps on my device with LAVA?
Scenario 1:
1. Flash images using fastboot
2. After android boots up or linux boots up, reboot and then stop at uboot
3. Do some tests in uboot.
Scenario 2:
1. Flash images using fastboot
2. After android boots up or linux boots up, interact with the device via uart (/dev/ttyUSB0) (suppose adb is disconnected)
3. Do some tests via uart.
Best regards,
Kingboli
Hi all,
I cancelled a test job but it still running on my device and showing testjob state as "cancelling". Is there any way to restart the device ,so that I can submit some other job .
Killing the specific job id on lava-master using lava-cli also not solving the problem.
Regards,
Smita Gumansingh
[cid:image001.jpg@01D2E5A8.A6033460]
Senior test Engineer, EPS PBU Software QA/Dev Ops
Mentor Graphics,
Nalapad Brigade Centre, Unit No:301 & 302
Mahadevapura, Bengaluru, Karnataka 560048, India
Office: +91 8067624338 |Cell: +91 9538349245
www.mentor.com<http://www.mentor.com/>
Hi all,
Our lava instance has been upgraded from 2019.03 to 2019.07 today.
Some of our teams need extra job context variables to be allowed, so I
added in /etc/lava-server/settings.conf the following parameter, as
described in one of your releases:
EXTRA_CONTEXT_VARIABLE: ["variable_1", "variable_2", "variable_3"]
It seems to work, but when I want to submit a job, sometimes I get the
warning "extra key not allowed" and if I hit the button "validate" many
times I finally get "valid definition". But sometimes I get "valid
definition" directly.
Is it a known issue?
Best regards,
Axel
