On 2/15/22 21:46, Maarten Dammers wrote:
Hi,
Why are we upgrading to Buster instead of Bullseye? According to https://wikitech.wikimedia.org/wiki/Operating_system_upgrade_policy Buster will be end of life around August this year. So we're either stuck with an older version for a while or we have to do this whole exercise again much sooner than we would like. Can you explain?
Hi there,
Legit question. I'm happy to elaborate:
* this was all discussed back in September 2021 in phabricator, see https://phabricator.wikimedia.org/T277653#7378774 and https://phabricator.wikimedia.org/T277653#7381146. Our conclusion was to don't skip Buster.
* we are hoping that there wont be a Buster->Bullseye migration for the grid. Hopefully by the time we need to remove Buster the Kubernetes backend will be 100% suitable solution for every tool.
* this migration work started before Debian Bullseye was released, with our intention being to complete it before the release. For a couple of reasons the project was delayed.
* in the grid case, the engineering effort to do a N+1 upgrade is lower than doing a N+2 upgrade. If we had tried a N+2 upgrade directly, things would have been much slower and difficult for us.
Your concern about doing the migration dance twice is 100% valid, and the only way to future-proof your tool is to remove dependency on GridEngine and migrate it to the Kubernetes backend.
regards.