Hi everyone,
I'm experiencing this error on setting up my U-Boot based device:
Unable to extract cpio archive '/var/lib/lava/dispatcher/tmp/17/extract-overlay-ramdisk-tv17mriz/ramdisk.cpio': Command '['cpio', '--extract', '--make-directories', '--unconditional', '--file', '/var/lib/lava/dispatcher/tmp/17/extract-overlay-ramdisk-tv17mriz/ramdisk.cpio']' returned non-zero exit status 2.
As documentation in https://docs.lavasoftware.org/lava/actions-deploy.html#deploy-action-roles states, the overlay will only be used, when a test action is defined in job - so I removed the test action (and the log states "[common] skipped lava-overlay - no test action.") but the error still appears.
For debugging I commented out the _cleanup() function in /usr/lib/python3/dist-packages/lava_dispatcher/job.py (and can see in job output that dispatcher tmp directory isn't cleaned anymore) but somehow the directory still disappears, so I don't know how to debug this further.
So what's going wrong here? How to get the overlay working?
Running LAVA 2023.01 on Debian 11.6, full log see attached.
Thanks in advance!
Hi Team,
Facing an issue while unpack the overlay in some custom location other than "/" directory. From the LAVA linaro documentation I came to know that the -C / command to tar is essential or the test shell will not be able to start. Below are the issues that I am facing because of this.
* My / directory is a read only one and I want it in some /etc or /var.
* The LAVA job is getting stuck while doing the unpack with the command below in job definition.
transfer_overlay:
download_command: cd /tmp; wget
unpack_command: tar -C /tmp -xzf
Below I am attaching the test log from LAVA.
[cid:image001.png@01D9455A.1AD8F6A0]
How we can solve the read only filesystem image issue for unpacking the overlay other than "/" directory ?
Regards,
Sarath P T
Hey there,
I set up LAVA, and ran some jobs (e.g qemu from tutorial) which went fine - but I noticed HTTP 500 status code on /api/help page. Which wasn't severe as I thought I might ask that later. But this monday I discovered that /accounts/login/ isn't working anymore, HTTP 500 status code too! Of course I didn't change anything... (everybody says that) and honestly can't say when this broke as I only logged in once after install. Django log says "django.contrib.sites.models.Site.DoesNotExist: Site matching query does not exist."and as far as I can see Sites reside in database?
As I'm new to LAVA I don't know how to debug and to solve this, can you please help me? Is there a way to re-install the sites/templates?
Logs that might be relevant are attached.
After install (and qemu setup) I noticed that /var is full. This is resolved (deleted some docker images), but might be relevant.
Running on current Debian 11 as apt install, recent LAVA version.
Thanks in advance
Stefan
Hey everyone,
in the announcement of LAVA 2023.01 it says:
> The support for Debian Buster has been dropped as Debian Buster does
> not provide support for the latest pyyaml versions.
So I updated my LAVA machines from buster to bullseye.
Now I have the problem that I cannot install lava-dispatcher anymore now due to the following error:
The following packages have unmet dependencies:
libbpf0 : Depends: linux-libc-dev (>= 5.14) but 5.10.162-1 is to be installed
libbpf0 seems to come from the LAVA repository:
libbpf0:
Installed: (none)
Candidate: 1:0.5.0-1~bpo11+1~lava1
Version table:
1:0.5.0-1~bpo11+1~lava1 500
500 http://apt.lavasoftware.org/release bullseye/main amd64 Packages
1:0.3-2 500
500 http://ftp.de.debian.org/debian bullseye/main amd64 Packages
While linux-libc-dev is part of the standard debian repositories:
linux-libc-dev:
Installed: 5.10.162-1
Candidate: 5.10.162-1
Version table:
*** 5.10.162-1 500
500 http://security.debian.org/debian-security bullseye-security/main amd64 Packages
100 /var/lib/dpkg/status
5.10.158-2 500
500 http://ftp.de.debian.org/debian bullseye/main amd64 Packages
This is the LAVA repository I am using:
deb http://apt.lavasoftware.org/release bullseye main
Is this a known issue? How do I correctly install LAVA 2023.01 on Debian Bullseye?
Thanks in advance and kind regards,
Tim
--
Tim Jaacks
SOFTWARE DEVELOPER
SECO Northern Europe GmbH
Schlachthofstrasse 20
21079 Hamburg
Germany
T: +49 40 791899-183
E: tim.jaacks(a)seco.com
Register: Amtsgericht Hamburg, HRB 148893 Represented by: Dirk Finstel, Marc-Michael Braun, Massimo Mauri
Hello Lava Users,
On deploying the latest LAVA release(LAVA 2022.11.1) packages after building the packages on Debian11 Host , we are facing issue with deploying lava-dispatcher-host package on Debian 11 host(11.6).
On checking further we noticed https://git.lavasoftware.org/lava/lava/-/blob/master/debian/control , lava-dispatcher-host package depends on base-files (<< 11.1) but base-files on system is 11.1+deb11u6.
To support Deploying on latest Debian bullseye release should this be updated or are we missing something.
$ sudo dpkg -i lava-dispatcher-host_2022.11.1+11+bullseye_all.deb
(Reading database ... 283093 files and directories currently installed.)
Preparing to unpack lava-dispatcher-host_2022.11.1+11+bullseye_all.deb ...
Unpacking lava-dispatcher-host (2022.11.1+11+bullseye) over (2021.10+10+buster) ...
dpkg: dependency problems prevent configuration of lava-dispatcher-host:
lava-dispatcher-host depends on base-files (<< 11.1) | python3-bpfcc (>= 0.21); however:
Version of base-files on system is 11.1+deb11u6.
Package python3-bpfcc is not installed.
lava-dispatcher-host depends on base-files (<< 11.1) | linux-headers-amd64 | linux-headers-arm64 | linux-headers-generic; however:
Version of base-files on system is 11.1+deb11u6.
....
Thanks,
Hemanth.
Hi All,
I'm looking to automate overall testing process and post the test results
in to JIRA instead of going to Lava CI.
Is there any way that I can use robot framework to run the tests on Lava
and so robot framework produces results in XML file that can be used to
post into issue tracker.
Any suggestions would be helpful
Thanks,
Pavan
Hello Team,
Good Day to All!
We are setting up a new device into LAVA Automation with the following
requirements.
Flash panel to R4 image
Copy delta to R5
Deploy R5
reboot
Make sure the panel boots to R5
The Test definition i'm using is as follows:
actions:
- deploy:
to: flasher
images:
package:
url:
https://artifactory.softwaretools.com/artifactory/mfgtools-***-0v3
- boot:
timeout:
minutes: 15
commands:
- boot
method: u-boot
prompts:
- 'root@hon-grip'
auto_login:
login_prompt: 'login: '
username: root
password_prompt: 'Password:'
password: root
login_commands:
- coredump --enable
- sysinfo
- ifconfig
- networkctl status
- wget http://192.100.**.**/oslat
- ostree static-delta apply-offline /home/root/oslat
- ostree admin deploy
35b3297cf3e4bc59d2a21e2ae9f7a02ef3f7a940e37389a7e9ae66a610c60b7
- reboot
After reboot command is executed, the panel again prompts for login and
password, whereas I cannot declare auto_login for the second time.
Please let me know how to disable auto login after the reboot (for second
time login)
Best Regards
*Pavan Kumar*
Hi Team,
I had a Query which is particularly to use lava-test-shell or other
binaries like lava-test-runner.
My board is booted with Linux and It has a POSIX environment but it doesn't
support either ssh/nfs due to the low memory footprint available and the
ethernet driver not fully functional.
To test/run my test-suite drivers How can I use lava-test-runner/
lava-test-shell ? Is it possible to test our suite using
lava-test-shell/runner where DUT doesn't have the ethernet/nfs support.
I am getting lava-test-shell timeout on the DUT console whereas Lava-worker
had all the binaries available by lava-overlay method.
Please find the attached test job definition/lava-job log files for your
reference. Kindly let me know the solution.
Hi teams,
I wonder if you could share an android cts lava job define to me, is it in https://validation.linaro.org/?
I want to have a reference, thanks!