re : Re: [Fase-developer] File Parser
Status: Planning
Brought to you by:
grevenx
|
From: <ma...@al...> - 2003-01-10 10:38:47
|
Alex: AFAIK once logfiles parsed then we could generate statistics, isn't this the point? Am I missing something? mazza. Posté par Alex le 10/1. I thought the whole point of parseing the log files was to get the data in a database. If it's not. Then what will anyone else be doing once that logfiles are parsed? On Fri, 10 Jan 2003, Fabio A Mazzarino wrote: > > KISS - Keep It Spetaculary Simple. > > There are other interpretations but it basicaly means *Keep It Simple*. > > File permission is not a problem, users always have some disk area, if > don't it means that they shouldn't be able to retrieve information about > apache logs. > > Fetching data from file only takes too long if files are too big, this > would be solved by log-rotating and consolidating logs, I think that > performance won't be a problem. > > Avoiding a database connection make the software much more avaiable for > most users, since a database manager is not required. > > Ow, I forgot to suggest another thing. Maybe a scretch of the whole > system and future modules would be interesting so we have something to > aim. Also a feature list for v0.1 and v0.2 should be interesting (don't > forget: release small, release often) > > Well, Even, I think that these two last task is about you, we can only > suggest. ;o) > > Waiting for your answer ;o) > > mazza > > > -------------------- > Looking for a better .sig. Suggestions are being accepted. > > > On Fri, 10 Jan 2003, [ISO-8859-1] Even André Fiskvik wrote: > > !On fredag, jan 10, 2003, at 01:09 Europe/Oslo, Fabio A Mazzarino wrote: > ! > !> > !> Hi folks: > !> > !> I'd like to make some suggestions. > !> > !> 1. I agree that it's better to use Apache::ParseLog. It's the Perl > !> way, > !> and it's a best pratice for programming: Lazyness > !> > !> 2. Is a database connectio *really* necessary? Maybe it's a good idea > !> to > !> rotate logs and consolidate them in a separete file, and then parse > !> each > !> time required. I mean, just like KISS ;o) > !> What do you think about it? > !> > !> mazza > ! > !What's KISS? And no, a database connection isn't really necessary, but > !users wouldn't have > !to worry about file permissions? When coming to think of it, it might > !not be too bad an idea to > !generate the statistics for a domain in it's own file. Then you > !wouldn't have to think about having > !a database server set-up, and we wouldn't need to think about the > !database abstraction to > !support them all. > !Anyone know what takes up most resources/takes most time, fetching the > !data from a database or a file? > ! > !- Even André > ! |