-
Type:
Bug
-
Status: Rejected
-
Priority:
Medium
-
Resolution: Won't Do
-
Affects Version/s: None
-
Fix Version/s: Kanban_Cleanup_2020_06_20
-
Labels:
I have been creating deliberately bad artifacts to test automatic rollback.
dd if=/dev/zero of=image.dat bs=1M count=0 seek=16
ARTIFACT_TOOL=/work/dmoseley/Mender/yocto/morty/build/beaglebone/tmp/sysroots/x86_64-linux/usr/bin/mender-artifact
${ARTIFACT_TOOL} write rootfs-image -t beaglebone -n mender-image-1.1 -u image.dat -o core-image-full-cmdline-beaglebone.mender
${ARTIFACT_TOOL} write rootfs-image -t vexpress-qemu -n mender-image-1.1 -u image.dat -o core-image-full-cmdline-vexpress-qemu.mender
${ARTIFACT_TOOL} write rootfs-image -t raspberrypi2 -n mender-image-1.1 -u image.dat -o core-image-full-cmdline-raspberrypi2.mender
This is based on the automated test script for this feature. When I try to load the artifact into my backend, I get a variety of errors:
- Artifact couldn't be uploaded. No size provided before the artifact part of the message or the size value is wrong.
- Artifact couldn't be uploaded. There was an error uploading the artifact
- Artifact couldn't be uploaded. internal error
Typically this will eventually succeed if I retry the upload but sometimes it takes 4-5 attempts.
I verified this behavior with the 1.0.x, master and git versions of the canned integration environment. I primarily used the mender-artifact tool that was built with my Yocto builds but also verified with the latest master branch version.
- relates to
-
MEN-2012 Explain cause when error uploading on Artifact
-
- Done
-