we need gmake here, added it to run dependencies
therefore updated PORTREVISION, otherwise you don't get
the png file...
- updated pkg-message. You need to configure SNMP on your cisco
or Juniper router
to make it more functional and userfriendly
- added cflowd's startup script to FILESDIR, add startup options to
it, that are only valid for the patched cflowd. So later we overwrite
the sample startup script of "stock" cflowd.
This is done this way, because cflowd also can be used without flowscan.
But *if* we use cflowd with flowscan, we have to manage things to work
with flowscan.
- install sample Makefile to graphs directory, so you can generate
graphs more easily as being described in the docu (see INSTALL)
- Added comments to pkg-message file, that this Makefile has to be
updated whenever you add or remove protocol types in
${PREFIX}/var/db/flows/bin/CampusIO.cf
- comment post-install section more..
- keep_me file could be removed, since we install the sample Makefile
now in the graphs subdir
- modify sample startup script
- run in background properly
- use variables for progs (grep, perl, kill, ...)
- style (use $var consistently, spaces after io redirection)
- install startup script as cflowd-flowscan.sh.sample
to get an order, cflowd ports startup script will be named
cflowd-base.sh.sample, so that cflowd will be started first.
this was overdue, since this complex port needs more docu,
explanation, preparation, etc ...
- installation directory moved
from ${PREFIX}/var/flows to ${PREFIX}/var/db/flows
to match our typical /var filesystem layout, but since I want
to package this port, I don't want to write into /var/filesystem
its up to the user, to move the data around later if he has a large
/var and likes the data to reside there.
- create needed directories in ${PREFIX}/var/db/flows (bin and graphs)
use a .keep_me file, so that things are packaged properly
- install needed documentation
the INSTALL file is essential as well as the various example file.
NOPORTDOCS only covers unneeded files, which are not crucial for
installation
- the ${PREFIX}/var/db/flows/bin directory carries the sample files
for the flowscan script. Therefore those sample files reside there.
Other sample files like the crontab settings are up to the user, so
they reside in ${PREFIX}/share/doc/flowscan
- install a sample rc file in ${PREFIX}/etc/rc.d, its deactivated, since
this is a port that needs thinking and handcrafting config files before
use...
- added a pkg-message file, that explains pre-requisites (cisco configuration)
and other post installation tasks as well as a BIG WARNING, that enabling
Netflow Switching and Netflow Data Export on your Cisco CORE or DISTRIBUTION
routers might affect the routers switching performance.
You should know IF you need to do it and if its o.k. IF you enable it.
- output pkg-message on port based installation as well since the warnings
are important.
- updated PLIST
- use LOCALBASE in BUILD dependency, so cflowd will be found
this should cure bentos build problems by multiply trying to
build/install cflowd (problem didn't show up here)
- bumped port revision
we also need to redefine WRKSRC because of this lowercase affair...
which I start to dislike, because if poeple learn about a software
named FlowScan, then they start looking for a port named FlowScan
and not flowscan ...
no portrevision bump necessary, since port didn't work properly.
now after giving the port a lowerrcase name you have to set the
DISTFILES variable right, so sources (which use upper/lowercase)
can be found.
PORTREVISION not bumped since port didn't work up to now,
so no new (working) functionality.
But anyway, thanks for doing the CVS work !
FlowScan is a tool to monitor and graph flow information from
Cisco and Riverstone routers in near real-time.
Amonst many other things, FlowScan can measure and graph traffic
for applications such as Napster.
A sample of what FlowScan can do is at: http://wwwstats.net.wisc.edu
WWW: http://net.doit.wisc.edu/~plonka/FlowScan/
