syslog threading model (maybe somewhat OT)
"Rainer Gerhards" <[email protected]>
| Newsgroups | gmane.comp.sysutils.loganalysis |
|---|---|
| Message-ID | <577465F99B41C842AAFBE9ED71E70ABA2786FB@grfint2.intern.adiscon.com> |
Hi list, this posting may be somewhat off-topic (some may think totally off-topic). I leave it to the wise list admin to decide if it still is covered by the list charter... I am the main author of rsyslog, an alternate GPLed syslogd. I have more or less completed the next major release and I am now thinking about the one after it. Obviously, there are a number of design decisions and one of the currently hot topics is the threading model. I have elaborated about what I have in mind here in my blog: http://rgerhards.blogspot.com/2007/08/syslog-worker-pools-future-hardwar e-and.html I know there are a lot of implementers of several syslog-related softwares (including syslogds) on this list. I would appreciate if you could have a look and drop me a comment or two. I think, however, that the overall topic could also be of interest for sysadmins, as such a discussion probably helps to better understand what's going on under the hood and what to watch out for. We had a similar situation quite a while ago where we talked about the SELP protocol (more or less syslog over tcp) on this list, and I think it was beneficial to the majority of readers. This is also the main reason why I try to post these off-topics things ;) Thanks, Rainer