Re: userlogging module?

From: Robert Collins <robert.collins@dont-contact.us>
Date: Fri, 16 Feb 2001 08:08:49 +1100

The MySQL acl patch provided here perviously might provide some inspiration, or even details :]

If you mean _just_ the actual write to disk bit, won't we have a terribly small helper?

I don't see any benefit in breaking any more of the access log code out to a helper - why not just have a queue of pending log
entries that are only cbLocked by the logging code. Then the logging code runs a pair of async i/o functions (similar to
clientsendmoredata & clientwritecomplete) that call the low level async functions

To add in a database we write a new set of low level async functions for that db, and configure them via squid.conf.

The reason I'm against the external helper is a) that the operation being effected is so trivial, and b) we already have most of an
async framework in squid - lets use it and c) the volume of access entries is going to very high for a single helper (and to keep it
serialised you can only have a single helper ?)

Rob

----- Original Message -----
From: "Adrian Chadd" <adrian@creative.net.au>
To: <squid-dev@squid-cache.org>
Sent: Friday, February 16, 2001 2:18 AM
Subject: userlogging module?

>
>
> Hi,
>
> I've had a feature request from someone.
>
> They'd like to be able to import the access.log info into an external
> database. Instead of doing it by tail -f'ing the logfile or
> periodically, they'd like to do it in real time.
>
> They'd also like to be able to log the username in the access.log entries
> when userauth is being used.
>
> Now, I'm sure the latter is easy. However, the former sounds like a nifty
> idea. I was thinking about breaking out the access.log code to use
> an external helper (to keep the data serialised).
>
> This has the added sideeffect of allowing us to implement async logging
> since the log file write is in a different process.
>
> What do people think?
>
>
> Adrian
>
> --
> Adrian Chadd "Romance novel?"
> <adrian@creative.net.au> "Girl Porn."
> - http://www.sinfest.net/d/20010202.html
>
Received on Thu Feb 15 2001 - 14:06:52 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:13:31 MST