lava-users

lava-users@lists.lavasoftware.org

1 participants
926 discussions

by Ilya Fedusiv

Hi, lava team! First question: I'm working with fastboot device, is it possible after test part put device again to fastboot mode, run some fastboot command ( not to flash ), boot again, run test. job steps: 1. deploy lxc 2. boot lxc 3. test lxc 4. deploy fastboot ( put to fastboot ) 5. boot fastboot device 6. run test 7. reboot and put device to fastboot mode 8. run fastboot commands 9. continue run test. Can you give any advice how to do it? Better to able reboot device inside test. I tried to reboot device from test part, lava just stuck and wait when reboot action will end. But looks like it does not recognize end. Second question: In job have lxc namespace and device namespace, how to run command_action from host os, not lxc? Best regards, Ilya Feduisv

5 years, 4 months

Use user commands

by Ilya Fedusiv

Hi, Lava Team! Want to use user_commands as written here : https://validation.linaro.org/static/docs/v2/actions-command.html Device dictionary has : {% set user_commands = {'switch_mode' : {'do': './usr/local/scripts/switch_mode.sh'}} %} Job https://pastebin.com/hhWbC207 I have lxc and custom script to set device flash-mode. But have error error_msg: 'common' is a reserved namespace that should not be present with other namespaces To fix it I added namespace to command: - command: namespace: tlxc name: switch_to_qdl But error again in validate before submit Invalid definition: extra keys not allowed @ data['actions'][3]['command']['namespace'] I have fastboot and custom flash methods. pre_power_command in deploy I use for fastboot. Need to use somehow custom script to switch to another mode, that's why I decided to use -command action. Please help how to fix error with namespaces or suggest another solution. Best regards, Ilya Fedusiv

5 years, 4 months

Using "command" actions in multinode jobs

by Ryan Harkin

Hi LAVA folks, I'm trying to create a multinode job based on iperf.yaml: https://github.com/Linaro/test-definitions/blob/master/automated/linux/iper… (note, this iperf.sh test doesn't actually work as intended, but that's a problem I'm debugging). The full job definition is at the bottom of this email. Anyway, I'm adding my own deploy and boot sections, as per working single node jobs. However, my deploy uses "command" actions, like this: - command: role: - host - guest namespace: dfu-firmware name: set_boot_to_nand These give error messages like this: Invalid definition: extra keys not allowed @ data['actions'][1]['command']['role'] And if I remove the "role" settings, I get a different error: Valid definition with warnings: Every action of a multinode job should have roles So it seems like I'm in a loop. Any suggestions?? Thanks, Ryan. ------------------------------------------------------ JOB DEFINITION ------------------------------------------------------ job_name: lces2-iperf timeouts: job: minutes: 60 action: minutes: 3 connection: minutes: 5 priority: medium visibility: public metadata: iperf-version: v3 protocols: lava-multinode: roles: host: device_type: rzn1d count: 1 timeout: minutes: 30 guest: device_type: rzn1d count: 1 timeout: minutes: 30 actions: #################### # Deploy QSPI images #################### - deploy: role: - host - guest namespace: dfu-firmware to: tmpfs images: sf_fsbl: image_arg: --alt sf_fsbl --download {sf_fsbl} url: https://snapshots.linaro.org/openembedded/schneider/linaro-warrior-4.9/rzn1… sf_trustedfw: image_arg: --alt sf_trustedfw --download {sf_trustedfw} url: https://snapshots.linaro.org/openembedded/schneider/linaro-warrior-4.9/rzn1… sf_uboot0: image_arg: --alt sf_uboot0 --download {sf_uboot0} url: https://snapshots.linaro.org/openembedded/schneider/linaro-warrior-4.9/rzn1… sf_uboot1: image_arg: --alt sf_uboot1 --download {sf_uboot1} url: https://snapshots.linaro.org/openembedded/schneider/linaro-warrior-4.9/rzn1… - command: namespace: dfu-firmware role: - host - guest name: set_boot_to_nand - boot: namespace: dfu-firmware role: - host - guest method: dfu timeout: minutes: 10 - command: namespace: test role: - host - guest name: set_boot_to_qspi ############### # TFTP/NFS ############### - deploy: namespace: test-nfs role: - host - guest timeout: minutes: 5 to: tftp os: oe kernel: url: https://snapshots.linaro.org/openembedded/schneider/linaro-warrior-4.9/rzn1… type: zimage dtb: url: https://snapshots.linaro.org/openembedded/schneider/linaro-warrior-4.9/rzn1… nfsrootfs: url: https://snapshots.linaro.org/openembedded/schneider/linaro-warrior-4.9/rzn1… compression: bz2 - boot: namespace: test-nfs role: - host - guest connection-namespace: dfu-firmware timeout: minutes: 20 method: u-boot commands: - setenv autoload no - dhcp - dhcp - env print - setenv serverip {SERVER_IP} - tftp {KERNEL_ADDR} {KERNEL} - tftp {DTB_ADDR} {DTB} - setenv bootargs 'console=ttyS0,115200n8 root=/dev/nfs rw nfsroot={NFS_SERVER_IP}:{NFSROOTFS},tcp,hard,intr,vers=3 rootwait coherent_pool=2M ip=dhcp' - '{BOOTX}' auto_login: login_prompt: 'login:' username: root prompts: - 'root@snarc-soca9:' - 'root@rzn1-snarc:' - test: namespace: test-nfs role: - host timeout: minutes: 30 definitions: - repository: metadata: format: Lava-Test Test Definition 1.0 name: run-iperf3-server description: "Run iperf3 in server mode" run: steps: - apt-get -y update - apt-get -y install iperf3 - if iperf3 -s -D; then lava-test-case run-iperf3-server --result 'pass'; else lava-test-raise 'Failed to run ipef3 server'; fi - lava-send server-ready ipaddr=$(lava-echo-ipv4 eth0) - lava-wait client-done - pkill iperf3 from: inline name: run-iperf3-server path: inline/run-iperf3-server.yaml - test: namespace: test-nfs role: - guest timeout: minutes: 30 definitions: - repository: https://git.linaro.org/people/chase.qi/test-definitions.git from: git path: automated/linux/iperf/iperf.yaml name: iperf3-test parameters: SERVER: lava-host-role # The MSG ID and KEY should be the same as the ones sent by lava-send above MSG_ID: server-ready MSG_KEY: ipaddr

5 years, 4 months

Re: [Lava-users] Unable to boot target via ipxe using nfs and tftp protocol via lava-docker container

by admirer mishra

Hi, I found the solution, we have to mount dispatcher.yaml file from lava-server in lava-master container as well beacuse lava-master is the one who communicate with DUT and run the job. regards admirer mishra

5 years, 4 months

Re: [Lava-users] Unable to boot target via ipxe using nfs and tftp protocol via lava-docker container

by admirer mishra

Hi Stevan, I am using latest lava-docker containers 2020.01 so using rest api I set the dispatcher ip and post it, which creates a dispatcher.yaml file with the content dispatcher_ip: '141.64.81.191' (not my ip) I checked it's there in the directory (/etc/lava-server/dispatcher.d/lava-dispatcher/dispatcher.yaml), but still no change in SERVER_IP or NFS_SERVER_IP (everything is same) while running the job, due to which not able to do tftp. I even tried with restarting the lava-server but still no change. Thanks for your time. regards, admirer mishra

5 years, 4 months

Unable to boot target via ipxe using nfs and tftp protocol via lava-docker container

by admirer mishra

Hi there, Yesterday, while going through the mail, I found about the docker-compose file provided by lava. The script is amazing it gives even more understanding of how actual lava-server and dispatcher work. I am a complete beginner just working on lava and docker from last 20 days. So, I am trying to boot a x86 device via ipxe using nfs and tftp protocol. In normal lava server running on host everything works fine because the host ip and target ip are on same local network so tftp and nfs works fine. But docker uses internal network and target is on local network so how to configure so that it works? I also used the docker compose file but the problem still persists. Is there any other configuration in the containers that I have to do make it work? Please help I am complete newbie in this field. Thanks for your time admirer mishra

5 years, 4 months

Lava Dispatcher Docker container tftp issue

by ROSHAN KUMAR

Hi, I am using the docker-compose file provided by lava community. I don't understand why there is a separate tftpd container and also how my device who is on host local network will communicate with it. Also same things I don't understand for nfs container. Please let me know Thanks and regards ROSHAN KUMAR

5 years, 4 months

Lava Server Docker container not working as Dispatcher

by ROSHAN KUMAR

Hi, I created a lava server docker container on amd64 machine using lavasoftware/lava-server docker image present on docker hub. But after running the container I didn't find any Dispatcher online except lava-logs. Did the current version of lava server docker container doesn't work as dispatcher similar to normal lava server. Please let me know. Thanks and regards ROSHAN KUMAR

5 years, 4 months

Mix device's type

by Ilya Fedusiv

Hi! I want to mix two flashing types for device : fastboot and custom ( run custom script ) I already did both types separately using deploy methods : fastboot and flasher How I can mix both types? And in web gui in device panel what does mean that error? Configuration Error: missing or invalid template. Jobs requesting this device type (mix-type) will not be able to start until a template is available on the master. Best regards, Ilya

5 years, 4 months

Dropping apache2 from lava-server container

by Remi Duraffort

Hello everyone, if you follow the latest developments, you might have seen that LAVA is now using whitenoise to allow gunicorn to serve static files. In the current releases (2020.01 and before), apache2 is forwarding dynamic requests to gunicorn and serving static files himself. In the next release, apache2 will forward all the requests to gunicorn that will serve both dynamic and static content. If you use the debian packages, apache2 will still be installed and configured automatically. We are proposing to drop the apache2 packages from the lava-server docker container. In fact, if you use the docker container, you have anyway to setup a reverse proxy to make the container visible to the outside world and to do the SSL termination. What do you think about this? Rgds -- Rémi Duraffort LAVA Architect Linaro

5 years, 4 months

← Newer
1
...
28
29
30
31
32
33
34
...
93
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

lava-users