[xenial] Test upgrade path: Ubuntu 14.04.5 → 14.04.5 with all updates → 16.04 #3965
Labels
Milestone
Comments
25 tasks
While running Even in the case of Fedora, we ask folks disable all third party repositories. |
|
The |
|
These are various User INPUT I had to give during the upgrade process.
|
|
Just a note that part of the task here is to do a system state comparison (installed packages, state of relevant config files, etc.) of the upgraded system with a fresh 16.04 install -- we want to make sure that upgrading from 14.04.5 to |
Packages in the default xenial app (based on libvirt_staging_xenial_app_staging):
Pakcages in upgraded xenial box from production app vm
Packages in the default xenial mon (based on libvirt_staging_xenial_mon_staging)
Pakcages in upgraded xenial box from production mon vm
Many of the upgraded package names came up as they are of new version and the version name is part of the package name. The default xenial image which we use in molecule also has a lot of extra packages, for example, The upgraded app vm gets |
|
While trying to upgrade the mon-server from the First,
Next, run
Then, it will ask to press
And after some time, it will give the details about upgrade and tell to press
Next, it will ask about restarting the services, I chose Then, it will ask about language selector, I have selected
Next, it will give information about
Then, the actual postfix configuration screen, I have selected no configuration required (default value), and pressed TAB to select
Next, the following input is required. I kept pressing Enter for the default input. After this it will show a few grub related screens and we will have to verify what we see on the actual hardware based installations. I did select any and kept pressing Enter.
The next screen is about removing extra packages, where I pressed
Press |
|
A new error I saw today: Seeing this again and again on my staging vm. @conorsch @eloquence @heartsucker @emkll @redshiftzero |
|
Good part is the above error is only happening on |
I suspect this means they moved the channel from |
|
I am still not sure why wifi-related packages are being installed in the upgrade scenario, per the list provided here: #3965 (comment) |
|
Ran through do-release-upgrade via Tails admin against Monitor Server on physical hardware (7-series NUC). Prompts were as follows:
Update completes successfully, can't reconnect after initial reboot (might just be impatience on my part), but after rebooting again I can connect via |
|
OK, just to recap next steps here based on the problems reported:
|
This was referenced Feb 20, 2019
|
Closing in favor of #4163 and #4164, which should be tackled as part of QA during this sprint. We'll open an issue for the error Kushal saw in staging if we encounter it again during testing. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Our plan of record is to upgrade from Ubuntu 14.04 to Ubuntu 16.04 by first ensuring that the system is updated to all latest packages available for Ubuntu 14.04 (not just security updates).
This is because the package version state of a SecureDrop server partially depends on whether or not the administrator has run the Ansible playbook, which causes a full package update.
To ensure that this strategy is viable, we should test the following upgrade path:
apt-get update && apt-get upgradedo-release-upgradeBeyond looking for showstopper bugs, we will want to compare the final system state with an install from a 16.04 base image to ensure that this upgrade does not result in odd inconsistencies that could cause problems later. This research may generate more tickets.
Part of #3204.
The text was updated successfully, but these errors were encountered: