Hi,
I am a bit confused about when mender commit should be called. I assume first mender install is used, then reboot (system reboots from updated partition) and them mender commit. But it appeared that reboot itself does not switch the partition. But install,commit,reboot does. Am I right?
Thank you
Mikhail
Assuming this is an embedded arm/yocto/uboot project, unless something changed in newer versions of mender client, historically mender -install in standalone mode was enough to do the update and switch the partitions in the bootloader environment. Then reboot which should startup in the newly updated partition, then do a mender -commit if happy with the newly booted partition.
If i recall correctly, use the following bootloader tool:
fw_printenv -n mender_boot_part
fw_printenv -n upgrade_available
before and after doing the mender -install and you should the bootloader env variables change
This is actually ubuntu x86_64 20 server, converted with mender-convert tool.
Yes I see that sequence is install - reboot -update.
BTW
~$ fw_printenv -n mender_boot_part
/usr/bin/fw_printenv: 1: eval: Bad substitution
but it works with no arguments.
~$ fw_printenv
bootcount=1
mender_boot_part=3
upgrade_available=0
mender_boot_part_hex=3
Thank you!
ok great, yer if i recall with the mender-convert on golden-images like ubuntu, the fw_printenv installed is not the same tool as the one that comes with uboot as I think it make changes to efi/grub instead on your device