Hi Cedric, On Wed, May 10, 2006 at 10:26:36AM +0200, Cédric Delaunay wrote:
> First : which machine should I use ? > I want collect about 9000 flows per minute. (this is a campus network). It > should be nice if I could collect, store data in a sql database and create > graphs on the same computer. The answer to the question greatly depends on the kind of aggregation you require rather the number of flows produced per minute by your network. If you, say, need to graph in VS out traffic per each of the 20k boxes .. you should be ok with a commodity P3-P4 collector. If you need to account in a SQL database EACH single micro-flow per EACH of the 20k boxes is different thing, instead. > The network represents more than 20000 machines and many sub-networks. I > need cacti to create graphs! Is there an other solution ( like create > rrd-tools graphs without cacti) to have a front-end showing statistics about > the network. Cacti creates RRD graphs offering bells and whistles: if you need a nice control panel, who can access which graph, arrange things hierarchically, etc. then Cacti is the way. Alternatively, you may choose the script way: a) get data (-s) each, say, 5 minutes from the memory table, push results in a file (output.txt) then erase (-e) table content: shell> pmacct -s > output.txt; pmacct -e b) a, say, Perl or shell/awk/sed script parses the produced output and generates graphs in a suitably arranged spool directory. It should be matter of very few lines of code. The scriptic way is, of course, available even going the SQL way aswell. As today pmacct is not featuring a RRD plugin, so there isn't support to write directly into RRD files. This bases on my very personal assumption (and if anyone reading has different ideas and/or proposals ... please let your comments to follow-up) that the collector should not care about creating/rotating files/directory trees (as this is likely to be prone to consistent delays and endless debates about the best-ever file hierarchy) and leave the work to external scripts/works. > Did somebody compile a paper like a tutorial which explain with details how > to deploy pmacct and nfacct collecting flows, create graphs and keep an > history for a year on a large network. I red pmacct.net documents but as a > newbie, I need more details. I think a detailed example could be useful. For examples about Cacti you can give a look to the step-by-step tutorial by Pedro Sanchez, available through pmacct homepage. To gather some first ideas about the scriptic way, you can take a look to Protograph; it is a working example written in shell. Cheers, Paolo _______________________________________________ pmacct-discussion mailing list http://www.pmacct.net/#mailinglists