Hi,

-----Original Message-----
From: cip-dev@lists.cip-project.org [mailto:cip-dev@lists.cip-project.org] On Behalf Of Jan Kiszka
Sent: Tuesday, August 25, 2020 2:59 PM
To: iwamatsu nobuhiro(岩松 信洋 □ＳＷＣ◯ＡＣＴ) <redacted>; cip-dev@lists.cip-project.org
Subject: Re: [cip-dev] [isar-cip-core][PATCH 1/2] ci: Rewrite using extends

On 25.08.20 07:33, nobuhiro1.iwamatsu@toshiba.co.jp wrote:

quoted

Hi,

quoted

-----Original Message-----
From: Jan Kiszka [mailto:jan.kiszka@siemens.com]
Sent: Saturday, August 22, 2020 1:45 AM
To: iwamatsu nobuhiro(岩松 信洋 □ＳＷＣ◯ＡＣＴ) <redacted>;

cip-dev@lists.cip-project.org

quoted

quoted

Subject: Re: [isar-cip-core][PATCH 1/2] ci: Rewrite using extends

On 21.08.20 17:04, Jan Kiszka wrote:

quoted

On 20.08.20 10:24, Nobuhiro Iwamatsu wrote:

quoted

Signed-off-by: Nobuhiro Iwamatsu <redacted>
---
 .gitlab-ci.yml             | 70 +++++++++++++++++++++++++++++---------
 scripts/deploy-cip-core.sh |  8 +++--
 2 files changed, 60 insertions(+), 18 deletions(-)

<snip>

quoted

quoted

quoted

-if [ -n "$DTB" ]; then
+if [ "$DTB" != "none" ]; then
 	aws s3 cp --no-progress

build/tmp/work/cip-core-*/linux-cip*/*/linux-cip-*/debian/linux-image-cip*/usr/lib/linux-image-*/$DTB
s3://download.cip-project.org/cip-core/$TARGET/

quoted

quoted

 fi

Unfortunately, this scale out to multiple jobs seem to cause download
issues, see e.g.

- https://gitlab.com/cip-project/cip-core/isar-cip-core/-/jobs/697844276
- https://gitlab.com/cip-project/cip-core/isar-cip-core/-/jobs/696946649

First thought it was a sporadic hic-up, but it reoccurs, now with
master. Is kernel.org throttling us here?

This may be a burden on kernel.org as snapshot reconstructs the image
on the server side. I'm not sure, when downloading images continuously,
access may be controlled by kernel.org.

So we need to avoid that for now, I guess. Back to single-job build, or
do we have an alternative download source?

OK, because I don't know any other good download source, I will change to single job.

quoted

quoted

And I'm afraid there is more broken, namely in the deployment that is
only triggered over master. Please have a look at the failing jobs.

I don't have no idea about this.
If we retry the test, it is successful and there may be other causes.

The wic image names changed, and that broke compression and uploading.

OK, I will fix this as well.

Jan

Best regards,
  Nobuhiro