Firmware demoted from qa.fwupd.org
Richard Hughes <[email protected]> Thu, 05 Jun 2025 08:10:36 +0000
| Newsgroups | dev.linux.lists.lvfs-announce |
|---|---|
| Message-ID | <RfB3jtYvh0D9SyN3h7LjfaZymqsPzmMGfzq60IinEEn4arUhSzCqfWwEDhthnRXp64ocflU80pgw_IaT5mfJknmQCoSNYkFjv8jgJOiQwus=@hughsie.com> |
Hi all, The LVFS staging "test" instance is used by us to verify things like databa= se upgrades and schema changes and contains a 3-year old static snapshot of= data to give a representative "lots of firmware, lots of reports, lots of = vendors, lots of clients" test case before we deploy onto the main producti= on LVFS. It's deployed in the same way, using the same containers onto the = same ECS machines with the same type of database -- so it's a great way for= us to test deploy things that might cause regressions. Yesterday an AI crawler started downloading all the files from the staging = instance of the LVFS (the one hosted at qa.fwupd.org) causing significant a= mounts of bandwidth to be used. The easiest thing to do was to just auto-de= mote all the test firmware from the public remotes -- which also made sense= on the logic that some vendors might not have been aware that the firmware= was also available there if they've since deleted/modified the old firmwar= e on the production LVFS instance. What I failed to do was to disable the e= mail notifications -- so some people who have had LVFS accounts for over 3 = years probably got quite a few emails about firmware "being demoted from qa= .fwupd.org". These emails can be ignored -- none of the firmware files on t= he production LVFS were touched. The QA instance now only has a few test fi= les visible which hopefully fixes the problem: https://qa.fwupd.org/lvfs/de= vices/ No action required; apologies for causing any alarm! Richard.