Hi
After receiving all routes from my BGP peer one of the CPU cores is constantly busy - 100% load by routing process
/system resource cpu print
# CPU LOAD IRQ DISK
0 cpu0 0% 0% 0%
1 cpu1 0% 0% 0%
2 cpu2 81% 0% 0%
3 cpu3 1% 1% 0%
4 cpu4 1% 0% 0%
5 cpu5 0% 0% 0%
6 cpu6 0% 0% 0%
7 cpu7 1% 1% 0%
8 cpu8 0% 0% 0%
/tool profile cpu=2
NAME CPU USAGE
networking 2 1.5%
management 2 3%
routing 2 90.5%
idle 2 5%
Is that normal behavior?
Hi
yes, thats normal, as BGP is running on one single core.
Regards
Patrick
Operations with large BGP tables are expensive. Anything that touches the table can cause high CPU usage (frequent BGP updates, routing table prints, etc.)
what can be benefit for using full bgp table ?
Multihoming with balancing between ISPs? Best Path Selection?
if use ospf for dnymic routing amd bgp full table can ccr1036 handle this ? and does it make performance for customer ? I have no idea about bgp full table so Im asking.
You really don’t need a full table unless you have multiple ISPs.
With a full table, your router can get two views of the entire internet and decide to use ISP1 to reach Google, and ISP2 to reach Facebook (for instance) based on how the two destinations look via both ISPs.
You can then go off the deep end with policy routing - BGP lets you get quite complicated - by making filter rules that tweak the routes’s metrics as you receive them.
Suppose you want to send traffic to Facebook via ISP3 while that link is up - then the in-filter on ISP3 can look for Facebook’s IP range (or in their case, the ASN would be a better choice) and matching routes have their local_pref increased higher than any other ISP’s version of the same prefix. This would make your entire network prefer ISP3 to reach Facebook (assuming that you are running iBGP between your routers - a VERY normal assumption)
I understand from here that I dont need full table…
and now here i create new question 
.. I use bgp betwen upstream provider and me on CCR1036 and other all network bridged. now i detect that when i ping 8.8.8.8 from bgp connected board which is ccr1036, ping is 40ms at the other board which is connectet this ccr1036 via sfp ping is 120ms. why its like that ? the problem can be that i dont use iBGP betwen routers ? i use ospf for dnymic routing also.
Latency is most likely not caused by your routing protocol.
OSPF is the correct protocol to use for your network’s internal routing. (Interior Gateway Protocol is the generic term for such a routing protocol - IGP)
iBGP is not intended to be used primarily as an IGP - its primary function is for your border routers to be able to share the external routing tables so that your network can make consistent decisions as a whole - so that router 7 knows you should reach such-and-such a destination using router 4, link 5…
Think of iBGP like choosing an airport to fly from when buying an airline ticket.
eBGP gives the list of flights / prices / times from a particular airport.
iBGP is like Travelocity which chooses the best of the flights based on some criteria.
OSPF is the “Google Maps” part that chooses the best route drive to the airport.
Anyway, the only way your routing protocol could be to blame is if it has chosen the wrong path - perhaps the shorter/faster path is down or else maybe it’s up but taking lots of errors or something like that.
Hi,
I understand that You offer me to use iBGP betwen routers. and it will give me better result with ospf.