On Mon, Dec 10, 2007 at 09:37:17AM +0100, Christoph Haas wrote: > The problem is not that I'm too lazy to maintain the package. It is just > not a useful replacement for other syslog daemons because the file names > that metalog writes to are not configurable. Yes, sometimes that can be a big problem. Log analyzers comes to mind. > Imagine you used the sysklogd and now replace it by metalog. Then > you will get a mess in /var/log/* because metalog can't write mail > log entries to e.g. /var/log/mail.log. System administrators should know the differences. > I have contacted the upstream about it but didn't get a reply so far. It seems that the original author lost interest in metalog. Last answer from him I have is two (or more) years old. > I don't intend to say that metalog is not a good piece of software. > But it doesn't fulfil the role of a "linux-kernel-log-daemon" (virtual > package). Risking to be boring and repeating over I must say it can be a good linux-kernel-log-daemon if sysadm knows what are the limitations and advantages (log files sizes are predictable). > Perhaps in the future metalog will be better configurable. Let me know > if you get somewhere with the upstream. I thought to take it (or creating a new project based on it) but I'm not sure how it can be useful to wider audience and don't want to waste time on project from which one person have benefit (me). But anyway, I'm working on patch which adds pid logging (original metalog doesn't log pids). Best regards
Attachment:
signature.asc
Description: Digital signature