Hi,
I'm learning to create a job to boot a mainline kernel on the pandaboard with debian rootfs in LAVA V2, but I keep running into these network related errors:
https://staging.validation.linaro.org/scheduler/job/168134#L625 https://staging.validation.linaro.org/scheduler/job/168134#L642
Here's my job definition: https://staging.validation.linaro.org/scheduler/job/168134/definition
Am I missing something to get the network working? I'd like to be able to run basic smoke tests as a start.
Regards, Lisa
On 13 March 2017 at 20:27, Lisa Nguyen lisa.nguyen@linaro.org wrote:
Hi,
I'm learning to create a job to boot a mainline kernel on the pandaboard with debian rootfs in LAVA V2, but I keep running into these network related errors:
https://staging.validation.linaro.org/scheduler/job/168134#L625 https://staging.validation.linaro.org/scheduler/job/168134#L642
Here's my job definition: https://staging.validation.linaro.org/scheduler/job/168134/definition
Am I missing something to get the network working? I'd like to be able to run basic smoke tests as a start.
That job is using an NFS rootfs which has "special" handlers for LAVA V1 master image support (lava-master-network) and it seems that those are getting in the way of the normal init process. There are some odd things with NFS on panda and mainline kernels. I'll see if I can get a working NFS rootfs later today and then I'll add it to the standard test jobs. In the meantime, are you able to do your tests with a ramdisk deployment?
On 14 March 2017 at 09:01, Neil Williams neil.williams@linaro.org wrote:
On 13 March 2017 at 20:27, Lisa Nguyen lisa.nguyen@linaro.org wrote:
Hi,
I'm learning to create a job to boot a mainline kernel on the pandaboard with debian rootfs in LAVA V2, but I keep running into these network related errors:
https://staging.validation.linaro.org/scheduler/job/168134#L625 https://staging.validation.linaro.org/scheduler/job/168134#L642
Here's my job definition: https://staging.validation.linaro.org/scheduler/job/168134/definition
Am I missing something to get the network working? I'd like to be able to run basic smoke tests as a start.
That job is using an NFS rootfs which has "special" handlers for LAVA V1 master image support (lava-master-network) and it seems that those are getting in the way of the normal init process. There are some odd things with NFS on panda and mainline kernels. I'll see if I can get a working NFS rootfs later today and then I'll add it to the standard test jobs. In the meantime, are you able to do your tests with a ramdisk deployment?
NFS on panda needs a rootdelay value on the kernel command line but we also need a change to the templates which will be this one:
https://review.linaro.org/#/c/17896/
Once that is merged, staging will update the following day.
NFS on panda needs a rootdelay value on the kernel command line but we also need a change to the templates which will be this one:
https://review.linaro.org/#/c/17896/
Once that is merged, staging will update the following day.
That's merged so you should be able to test it tomorrow morning on staging.
On 14 March 2017 at 13:13, Remi Duraffort remi.duraffort@linaro.org wrote:
NFS on panda needs a rootdelay value on the kernel command line but we also need a change to the templates which will be this one:
https://review.linaro.org/#/c/17896/
Once that is merged, staging will update the following day.
That's merged so you should be able to test it tomorrow morning on staging.
Reference job: https://staging.validation.linaro.org/scheduler/job/168232 https://staging.validation.linaro.org/scheduler/job/168232/definition
On 15 March 2017 at 00:26, Neil Williams neil.williams@linaro.org wrote:
On 14 March 2017 at 13:13, Remi Duraffort remi.duraffort@linaro.org wrote:
NFS on panda needs a rootdelay value on the kernel command line but we also need a change to the templates which will be this one:
https://review.linaro.org/#/c/17896/
Once that is merged, staging will update the following day.
That's merged so you should be able to test it tomorrow morning on staging.
Reference job: https://staging.validation.linaro.org/scheduler/job/168232 https://staging.validation.linaro.org/scheduler/job/168232/definition
The reference job worked for me. Thank you!
Is there any way that I can use a developer debian rootfs from releases.linaro.org[1] without running into an init issue[2]? The developer rootfs usually has all the build packages I need to create binaries for tests such as ebizzy[3] on the target device.
[1] http://releases.linaro.org/debian/images/developer-armhf/ [2] https://staging.validation.linaro.org/scheduler/job/168297#L769 [3] http://git.linaro.org/qa/test-definitions.git/tree/ubuntu/ebizzy.yaml
--
Neil Williams
neil.williams@linaro.org http://www.linux.codehelp.co.uk/ _______________________________________________ Lava-users mailing list Lava-users@lists.linaro.org https://lists.linaro.org/mailman/listinfo/lava-users
Allow me to wade in here.
Lets go back to the beginning.
I think what everyone has missed here, is that in the first job definition you showed: https://staging.validation.linaro.org/scheduler/job/168134/definition has hardcoded boot commands in the job. If you look at them, they set "root=/dev/mmcblk0p2" which is using a rootfs from the MMC! By pure luck that happens to have a lava master image on it. But you're not actually booting the nfsrootfs defined in the job at all.
So I took your original definition, removed the hard coded boot commands (and replaced with "nfs" to use LAVAs built in ones), fixed a few problems such as the prompt being wrong, updated your kernel/dtb urls, and here it is booted fine. https://staging.validation.linaro.org/scheduler/job/168316 https://staging.validation.linaro.org/scheduler/job/168316/definition
Overriding boot commands is a bad idea in general. If the LAVA ones aren't doing what you need, talk to the lab to update them.
Matt.
On 16 March 2017 at 01:08, Lisa Nguyen lisa.nguyen@linaro.org wrote:
On 15 March 2017 at 00:26, Neil Williams neil.williams@linaro.org wrote:
On 14 March 2017 at 13:13, Remi Duraffort remi.duraffort@linaro.org wrote:
NFS on panda needs a rootdelay value on the kernel command line but we also need a change to the templates which will be this one:
https://review.linaro.org/#/c/17896/
Once that is merged, staging will update the following day.
That's merged so you should be able to test it tomorrow morning on staging.
Reference job: https://staging.validation.linaro.org/scheduler/job/168232 https://staging.validation.linaro.org/scheduler/job/168232/definition
The reference job worked for me. Thank you!
Is there any way that I can use a developer debian rootfs from releases.linaro.org[1] without running into an init issue[2]? The developer rootfs usually has all the build packages I need to create binaries for tests such as ebizzy[3] on the target device.
[1] http://releases.linaro.org/debian/images/developer-armhf/ [2] https://staging.validation.linaro.org/scheduler/job/168297#L769 [3] http://git.linaro.org/qa/test-definitions.git/tree/ubuntu/ebizzy.yaml
--
Neil Williams
neil.williams@linaro.org http://www.linux.codehelp.co.uk/ _______________________________________________ Lava-users mailing list Lava-users@lists.linaro.org https://lists.linaro.org/mailman/listinfo/lava-users
Lava-users mailing list Lava-users@lists.linaro.org https://lists.linaro.org/mailman/listinfo/lava-users
On 16 March 2017 at 02:50, Matt Hart matthew.hart@linaro.org wrote:
Allow me to wade in here.
Lets go back to the beginning.
I think what everyone has missed here, is that in the first job definition you showed: https://staging.validation.linaro.org/scheduler/job/168134/definition has hardcoded boot commands in the job. If you look at them, they set "root=/dev/mmcblk0p2" which is using a rootfs from the MMC! By pure luck that happens to have a lava master image on it. But you're not actually booting the nfsrootfs defined in the job at all.
So I took your original definition, removed the hard coded boot commands (and replaced with "nfs" to use LAVAs built in ones), fixed a few problems such as the prompt being wrong, updated your kernel/dtb urls, and here it is booted fine. https://staging.validation.linaro.org/scheduler/job/168316 https://staging.validation.linaro.org/scheduler/job/168316/definition
Overriding boot commands is a bad idea in general. If the LAVA ones aren't doing what you need, talk to the lab to update them.
Thanks for your input, Matt. Yeah I wouldn't normally override boot commands, but when LAVA complained about a missing 'commands' field in my job definition, it wasn't obvious that I could have typed 'nfs' or 'ramdisk-tftp'.
Occasionally I may need to pass in an extra argument to the kernel command line to test a colleague's patchset, but I believe Neil's change takes care of that.
Matt.
On 16 March 2017 at 01:08, Lisa Nguyen lisa.nguyen@linaro.org wrote:
On 15 March 2017 at 00:26, Neil Williams neil.williams@linaro.org wrote:
On 14 March 2017 at 13:13, Remi Duraffort remi.duraffort@linaro.org wrote:
NFS on panda needs a rootdelay value on the kernel command line but we also need a change to the templates which will be this one:
https://review.linaro.org/#/c/17896/
Once that is merged, staging will update the following day.
That's merged so you should be able to test it tomorrow morning on staging.
Reference job: https://staging.validation.linaro.org/scheduler/job/168232 https://staging.validation.linaro.org/scheduler/job/168232/definition
The reference job worked for me. Thank you!
Is there any way that I can use a developer debian rootfs from releases.linaro.org[1] without running into an init issue[2]? The developer rootfs usually has all the build packages I need to create binaries for tests such as ebizzy[3] on the target device.
[1] http://releases.linaro.org/debian/images/developer-armhf/ [2] https://staging.validation.linaro.org/scheduler/job/168297#L769 [3] http://git.linaro.org/qa/test-definitions.git/tree/ubuntu/ebizzy.yaml
--
Neil Williams
neil.williams@linaro.org http://www.linux.codehelp.co.uk/ _______________________________________________ Lava-users mailing list Lava-users@lists.linaro.org https://lists.linaro.org/mailman/listinfo/lava-users
Lava-users mailing list Lava-users@lists.linaro.org https://lists.linaro.org/mailman/listinfo/lava-users
lava-users@lists.lavasoftware.org
-
Lisa Nguyen -
Matt Hart -
Neil Williams -
Remi Duraffort