Hello, setting fallback mirrors can help when there are mirror problems. Adding an option to detect and use the fastest mirror automatically can be helpful too.
Hi thanks proposing this. Cauldron is not intended to be used as daily basis, nor in production environnent. This has been previously discussed on Dev Mailing List. Mirrors can go stale, or being out-ohttps://mirrors.mageia.org/statusf-sync for several moment. You can check here: https://mirrors.mageia.org/status We can't have the infrastructure for having "fallback" mirror. Other solution when something goes wrong: # urpmi --clean which remove wrong or corrupted downloaded RPM that prevent urpmi to install New packages. If mirror has a wrong synthesis hdlist file, this is due to incomplete synchronization with main mirrors. Solution: try again 1 hour after. So, closing this WONTFIX.
CC: (none) => ouaurelienStatus: NEW => RESOLVEDResolution: (none) => WONTFIX
I think the suggestion is not about infrastructure, but a client side solution. The problem is not limited to cauldron, sometimes users or update testers have some problems. Enhancements requests are set to Cauldron even when the problem is in current releases. The messages the user see as result could indeed be improved, and offering the actions proposed in comment 1. As it is now, the messages the user see - conflicts etc - are confusing for them and hide the actual problem when it is aused by partiially updated mirror. Add to that the user may need to issue the clean command to get out of it. That said, the mirror i use http://ftp.acc.umu.se/mirror/mageia/ is very reliable. But I cant decide about reopening this bug, it is a well known and discussed problem anyway. Problem is lack of resources :(
CC: (none) => fri