Hide Forgot
Description of problem: Now that we have systemd journal is there any reason for us to still have rsyslog in core? Spins that want to provide their target user base with enterprise feature full file logging solution would of course have add syslog-ng or rsyslog to their .ks file and users manually install either one of those two components if they want/need it. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Minimal install should still have a log daemon of some sort, and realistically there are more packages that expect its presence that don't actually have RPM Requires on it. Those intending a really really really minimal system can remove it where necessary... in the grand scheme of things, they likely don't want authconfig, openssh, or possibly even a package manager, all things which we ship in @core.
(In reply to comment #1) > Minimal install should still have a log daemon of some sort, Are you saying that systemd journal is inadequate for that part? If so then this decision makes absolute sense and journald should be disabled by default. > and realistically > there are more packages that expect its presence that don't actually have RPM > Requires on it. Yeah that's the penalty you get for not having proper guidelines but hey instead of fixing let's just keep things as broken as they are in that regard...
(In reply to comment #2) > (In reply to comment #1) > > Minimal install should still have a log daemon of some sort, > > Are you saying that systemd journal is inadequate for that part? It's inadequate for the expectations of the users and administrators at this time, especially since it's only existed for a month or two.
(In reply to comment #3) > (In reply to comment #2) > > (In reply to comment #1) > > > Minimal install should still have a log daemon of some sort, > > > > Are you saying that systemd journal is inadequate for that part? > > It's inadequate for the expectations of the users and administrators at this > time, especially since it's only existed for a month or two. It has existed since systemd v38 I believe rather then v37 btw and this proposal was against rawhide(F18) not F17 which should given both users and administrators at least additional 6 months to toy with it once F17 had been released and any outstanding issues against it being fixed. It also would give us time to fix packages using the same approach we used in F16 with migrating legacy sysv init scripts. Granted that my changes to the FPG would have been approved by FPC first for that to happen which I have now asked for that draft to be dropped since it serves no purpose since rsyslog was not remove from core. But it's interesting to see that some technology changes are fine to be imposed to users and administrators within in one release cycles ( like systemd it self ) while others are not.