it's been a 2 week testing period of MPD5 , and it works like a charm .. we are moving to production with MPD5 as an LNS to terminate DSL users .. however , some questions arose :
1-how to get an SNMP MIB to show the number of connected users - virtual interfaces -
2- Bandwidth utilisation for each interface , and collectively .
3- i've read somewhere that SNMPD could cause the server to crash .. is this issue solved yet !?!?
4- what's the deal with compiling MPD5 with or without the FLOWTABLE .. !?!? is that a serious issue !?
thanks in advance for your quick reply
For user level accounting,we rely on RADIUS records. For more detailed forensic analysis netflow on the outside facing interface is meets our needs as we can tie the IP traffic back to the individual user. (we actually use argus, but its essentially the same thing). snmpd wont cause crashes, but it adds a lot of CPU overhead. I forget which is worse- bsnmp or netsnmp in terms of how its implemented. With either of those, you can extend the enterprise mib to ask mpd which can spit out the number of users from the command line. e.g. on bsnmp, I have the following so I can check the temp of the CPUs
extNames.2 = "temp0"
extCommand.2 = "/sbin/sysctl -n dev.cpu.0.temperature | /usr/bin/sed -e 's/C//g'"
extNames.3 = "temp1"
extCommand.3 = "/sbin/sysctl -n dev.cpu.1.temperature | /usr/bin/sed -e 's/C//g'"
OID .126.96.36.199.4.1.2021.8.1.101.2 and .3
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.