On Dec 24, 2018 at 09:09 PM Sarah Julia Kriesch wrote:
It is always hard to guess. There is a guess "when no problems appear". But even if the upgrade is always tested on the staging machine, things don't go always as expected.
That sounds like you have to test with a "fresh" VM. Perhaps you can receive the same behavior if you will ask after a snapshot of the live system for test upgrades.
Yes, I used a fresh image. Some time ago, production instance was installed from the same data set, but production setup went through the set of incremental upgrades. Starting with a whole filesystem snapshot could be a good idea. Or even making both and evaluating differences. By the way, I can clean-up the staging scripts a bit, and provide Weblate easy installation tool to others.
In addition, you have to use a separate database with the database dump of the live system (like now). Then the ip address of the snapshot vm and the database configuration have to be changed. After that you should be able to reproduce all better.
Yes, I do that. In addition, I have another VM that simulates openSUSE infrastructure: haproxy server with NPN HTTPS setup and the database server. But even with the best ever setup, problems can appear. This time I discovered a race condition that never appeared on the production machine, but appeared on the staging machine. https://github.com/WeblateOrg/weblate/issues/2467 -- Best Regards / S pozdravem, Stanislav Brabec software developer --------------------------------------------------------------------- SUSE LINUX, s. r. o. e-mail: sbrabec@suse.com Křižíkova 148/34 (Corso IIa) tel: +49 911 7405384547 186 00 Praha 8-Karlín fax: +420 284 084 001 Czech Republic http://www.suse.cz/ PGP: 830B 40D5 9E05 35D8 5E27 6FA3 717C 209F A04F CD76 -- To unsubscribe, e-mail: opensuse-translation+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-translation+owner@opensuse.org