Hi,
I’m currently testing the workflow for my first iOT devices that is using a 4G connection. I have a raspbian image where I installed al the needed packages.
I did the following:
-
installed self hosted mender on Amazon AWS EC2 server with 2gb memory and 4gb swap 30 gb storage with success
-
dd
copied the image -
used
docker-mender-convert
with following settings./docker-mender-convert from-raw-disk-image
–storage-total-size-mb 29000
–raw-disk-image $RAW_DISK_IMAGE
–mender-disk-image $MENDER_DISK_IMAGE
–device-type $DEVICE_TYPE
–artifact-name ARTIFACT_NAME \ --bootloader-toolchain arm-buildroot-linux-gnueabihf \ --server-cert (pwd)/server.crt
–server-url “https://url-of-server”
I added storage-total-size
with max of sdcard.
I got an working sdimage
I wanted to test by copying the .mender
artifact to the server. The file is 1.31 gigabyte.
I got the device registered in the Mender dashboard. So that’s working.
I tried again with a .mender file of less than 100 mb with the same issue. Checked if there is enough space on the server.
After that I want to make an update of the mender sdimage. I changed some code on the mender sdimage and used dd
to make a copy of the sd card again. So it’s not the golden image, but the mender sdimage created by mender-convert
.
I have a few questions:
- Should I use the mender sdimg or the original image for update?
- How does Mender update the rpi? The file is 1.31 gb. When I create a new one with mender convert, I would expect that the file size would be the same. I think 1.31gb is way to large to send over 4G connection to the devices. The difference between prev image and updated one is max a few megabytes.