Gehel added a comment.
In https://phabricator.wikimedia.org/T133026#2221842, @daniel wrote: > @Smalyshev yes, that's what I suspected. But that's not how load balancing for static resources should work. It screws with the logic of IFM headers. Varnish should hide the differences, there should be only one state of the virtual resource represented by a url. > > Basically, this kind of load balancing causes more load, because static resources that are already cached in the browser are often delivered again, when the request is hitting a different box, where the resources has a different etag. > > The simplest fix would probably be to just force the mtime on all the static files. Forcing mtime would need to be done by Trebuchet during deployment. Varnish does not have shared state as far as I know, so it will not be able to hide the mtime differences. While it might be possible to do it via trebuchet, I would prefer to remove etags, push for a very large caching time and have a clean cache busting solution (https://gerrit.wikimedia.org/r/284428). TASK DETAIL https://phabricator.wikimedia.org/T133026 EMAIL PREFERENCES https://phabricator.wikimedia.org/settings/panel/emailpreferences/ To: Smalyshev, Gehel Cc: gerritbot, daniel, TerraCodes, Gehel, Lydia_Pintscher, JanZerebecki, Jonas, Aklapper, Avner, Lewizho99, Maathavan, debt, D3r1ck01, FloNight, Izno, jkroll, Smalyshev, Wikidata-bugs, Jdouglas, aude, Deskana, Manybubbles, Mbch331 _______________________________________________ Wikidata-bugs mailing list Wikidata-bugs@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs