On Sat, Mar 08, 2014 at 10:40:27AM -0600, Stephen Martin wrote:
If you want to use arch in a server environment, you should probably use your own repo to be safe.
Have a testing machine be on rolling release. When that machine is stable, push its packages to a private repo and update your server via the private repo.
That is, IMNSHO, a terribly sensible idea for reliable running of critical installations, and I've been implementing this procedure for a while now with good results: Manually, with a lot of pacman ad-hackery. :P I have yet to find some "glue" to make this easier and more fool proof for production usage. Is it possible that really no-one has yet created a script or Makefile or anything to automate the process of: 1) Updating a running test installation, including AUR and custom packages, to a specific point in time (most often: now) 1a) Alternative/Prerequisite to 1): Create a testing repository containing exactly those packages present after an update of a (hypothetical) testing system, i.e. a given set of packages. 2) Snapshotting the package state of a given running installation (or offline root), including AUR and custom packages 3) Synchronizing a specific custom repository to this exact package state determined in 2), without losing the custom and AUR packages on the way, and without keeping old/replaced packages around needlessly. Doing this is easy, superficially, but *practically* there are a few pitfalls involved, which kept me from just hacking up some scripts "real quickly" myself, dooming me to a life of repeating error-prone steps manually. Does anyone know of an existing and reliable solution to this administrative problem, or is this something to contribute back one day? Thanks, Dennis -- "Den Rechtsstaat macht aus, dass Unschuldige wieder frei kommen." Dr. Wolfgang Schäuble, Bundesinnenminister (14.10.08, TAZ-Interview) 0D21BE6C - F3DC D064 BB88 5162 56BE 730F 5471 3881 0D21 BE6C