On Wed, 30 Jun 2010, Sean Conner wrote:
> It was thus said that the Great Rainer Gerhards once stated:
>>
>> Note that being an evolution, any currently existing rsyslog.conf would also
>> be a valid new conf (via the *same* parser). I have not thought out the full
>> semantics and thousand other things that need to be thought of -- this
>> actually opened up a can of worms ;) However, I find the proposed format
>> seems to be a good compromise between the need to preserve and the need to
>> move on, and between the need for simplicity and the need for power.
>>
>> As such, before I invest even more time into that format, I'd like to get
>> some feedback on what you folks think ;)
>
> You need to be very careful when creating your own language. I did that
> once (created my own langauge, back in college, just for fun) and recently
> benchmarked it against commonly used scripting languages today and the
> results were horrible---mine was the slowest tested (Perl, Lua, Python; I
> was afraid of testing Ruby in fear to losing there too). That's an issue to
> watch out for---a slow interpreter (or particularly bad internal code
> representation).
one key thing is that this language is not a script to be executed as logs
come in, it's a config file that will get read in to configure rsyslog
internals (compiiled into the rsyslog internal function calls effectivly)
no off-the-shelf language is fast enough for the data rates that rsyslog
is able to reach today.
David Lang
> Sure, now you're just adding an "if ... then ... [else ...]" construct
> now, but as people get used to this, there might be calls for additional
> functionality, like variables (why can't I declare file paths as constants
> in one location, etc etc) and pretty soon you'll have a fully blown
> programming language (maybe not this year, maybe not next, but some day).
> That's a second issue to watch out for---feeping creaturism.
>
> Other than that, I don't really have much to say about the syntax. I know
> you're trying to work within an existing framework so you do have some
> contraints and I've certainly seen (and worked) with worse configuration
> files (sendmail.cf comes to mind).
>
> One last thing though, in looking over your proposed syntax for
> RainerScript I see that's it already is awfully close to Lua:
>
> ruleset remote10515 {
> if pri("mail.*") then {
> action(type="omfile" file="/var/log/remote10514")
> action(use="dynfile")
> action(type="udpfwd" action.execonlyonce="5sec"
> target="192.168.1.2" port="514")
> }
> action(type="udpfwd" target="192.168.1.3"
> action.previousfailed="on")
> action(type="omfile" file="/var/log/catchall")
> if $severity == 'error' and $msg contains 'Link 2' then
> action(type="ommail" server="192.168.1.3"
> from="[email protected]"
> to="[email protected]"
> subject="###error \"detected\"###")
> }
>
> could be translated into Lua as:
>
> function remote10515()
> if pri("mail.*") then
> action { type="omfile" , file="/var/log/remote10514" }
> action { use="dynfile" }
> action {
> type = "udpfwd" ,
> action = { execonlyonce="5sec" },
> target = "192.168.1.2",
> port = 514
> }
> end
>
> action {
> type="udpfwd" ,
> target="192.168.1.3" ,
> action = { previousfailed = true }
> }
> action { type="omfile" , file="/var/log/catchall" }
> if severity == 'error' and string.find(msg,'Link 2') then
> action {
> type="ommail",
> server="192.168.1.3",
> from="[email protected]",
> to="[email protected]",
> subject=[[###error "detected"###]]
> }
> end
> end
>
> It's a shame you're rejecting Lua---it's considered one of the fastested
> scripting languages and dead simple to embed. If you would like to see what
> could be done with Lua, you might want to check out my own syslogd I wrote
> that embeds Lua:
>
> http://www.conman.org/software/syslogintr/
>
> There is a speed hit (about 35%) in using Lua vs. straight C, but so far,
> speed hasn't been an issue in what I do.
>
> -spc (which includes the processing of multiple thin logs into one fat
> log)
>
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog
> http://www.rsyslog.com
>
_______________________________________________
rsyslog mailing list
http://lists.adiscon.net/mailman/listinfo/rsyslog
http://www.rsyslog.com
