You can also do v9, but that comes with a load of restrictions. Better to
do ipfix IMO. L2 info is only grabbed off of L2 interfaces if I remember
correctly.
Dan
https://www.juniper.net/documentation/en_US/junos/topics/concept/inline-sampling-overview.html
On Fri, Dec 29, 2017 at 1:01 PM
Hello all you wonderful JUNOS geeks :)
Happy New Year!
Couple of quick questions:
Current platform
MX480
RE-1800x4
MPC3-3D
MPC2-3D
1Gig and 10Gig MIC's
SCBE
Wanting to get flow data for both IPv4 and IPv6. Seems I need IPFIX
for this
I'm also trying to get MAC addresses into my flows so
We have similar traffic-levels (a bit more actually) and sample 1:100
and handle the analysis on an E3 3.0Ghz/16GB RAM/2x500GB SATA SW-RAID1
with 10G card with no problems and loads of capacity to spare.
For those traffic levels you certainly don't need a dedicated 10G card
for the netflow. 1G
That's JFLOW, and as sthaug pointed out, I don't need 10G at all, it
barely hits 3% utilization on the NIC's during peaks.
//T
On Fri, Mar 28, 2014 at 9:14 AM, Gavin Henry ghe...@suretec.co.uk wrote:
On 28 Mar 2014 08:10, Timh Bergström timh.bergst...@videoplaza.com
wrote:
We have similar
* Andrew Jones a...@jonesy.com.au [2012-08-12 05:32]:
Hi All,
There is a behaviour in jflow v9 which is frustrating me whereby once
flows reach the active timeout the flow is exported, the packet and byte
counters reset, but the flow remains in the flow table and the start time
is not
Dear community member,
We are producing a rotator with 02 MX480-logical systems (LS1 and LS2) and
routing-instance-virtual routers (ls1 and ls2), I need to set up
collection JFLOW
(samping).
logical-systems {
LS1 {
interfaces {
ge-0/0/0 {
unit 0 {
6 matches
Mail list logo