Hi Osama,

Very good to hear from you. About the UDP ports: field types #180 and
#181 are supported but the fact the template contains both these two
and their traditional counterparts, field types #7 and #11, kills the
heuristics currently in place to decide which one to pick. What is the
reason behind putting all together in a single template? If it makes
sense to you to remove such redundancy and have pmacct working, good;
otherwise please send me privately a brief capture so to improve the
current heuristics.

About the MPLS label fields: these are not supported (well, there is
an indirect way to aggregate on them via pre-tagging, but i will not
propose it to you as a viable solution). I more wonder what is the
use-case behind your interest into them - it would be great if you
can elaborate on it.

Cheers,
Paolo

On Sun, May 05, 2013 at 11:38:06AM +0200, Osama Abuelsorour wrote:
> Hello,
> 
> I've started using pmacct as a collector for few Cisco routers. We are very 
> interested in mpls labels and other field types. pmacct did pick up most of 
> them but not all. In addition, it didn't pick up any L4 information because 
> they were received on the more specific types such as udpSourcePort instead 
> of l4 src port, etc.
> Here is the NFv9 template:
> 
> DEBUG ( default/core ): NfV9 template type : flow
> DEBUG ( default/core ): NfV9 template ID   : 258
> DEBUG ( default/core ): -----------------------------------------------------
> DEBUG ( default/core ): |    pen     |     field type     | offset |  size  |
> DEBUG ( default/core ): | 0          | mpls label 1       |      0 |      3 |
> DEBUG ( default/core ): | 0          | mpls label 2       |      3 |      3 |
> DEBUG ( default/core ): | 0          | IPv4 src addr      |      6 |      4 |
> DEBUG ( default/core ): | 0          | IPv4 dst addr      |     10 |      4 |
> DEBUG ( default/core ): | 0          | L4 src port        |     14 |      2 |
> DEBUG ( default/core ): | 0          | L4 dst port        |     16 |      2 |
> DEBUG ( default/core ): | 0          | 186                |     18 |      2 |
> DEBUG ( default/core ): | 0          | 180                |     20 |      2 |
> DEBUG ( default/core ): | 0          | 181                |     22 |      2 |
> DEBUG ( default/core ): | 0          | 234                |     24 |      4 |
> DEBUG ( default/core ): | 0          | 44                 |     28 |      4 |
> DEBUG ( default/core ): | 0          | input snmp         |     32 |      4 |
> DEBUG ( default/core ): | 0          | 89                 |     36 |      1 |
> DEBUG ( default/core ): | 0          | BGP IPv4 next hop  |     37 |      4 |
> DEBUG ( default/core ): | 0          | 45                 |     41 |      4 |
> DEBUG ( default/core ): | 0          | output snmp        |     45 |      4 |
> DEBUG ( default/core ): | 0          | direction          |     49 |      1 |
> DEBUG ( default/core ): | 0          | 235                |     50 |      4 |
> DEBUG ( default/core ): | 0          | 95                 |     54 |      4 |
> DEBUG ( default/core ): | 0          | in bytes           |     58 |      8 |
> DEBUG ( default/core ): | 0          | first switched     |     66 |      4 |
> DEBUG ( default/core ): | 0          | last switched      |     70 |      4 |
> DEBUG ( default/core ): -----------------------------------------------------
> DEBUG ( default/core ): Netflow V9/IPFIX record size : 74
> 
> Here is a sample output of nfacctd:
> TAG         TAG2        CLASS             VLAN   ETYPE  SRC_IP           
> DST_IP           SRC_PORT  DST_PORT  PROTOCOL    TOS    PACKETS     FLOWS     
>   BYTES
> 0           0           unknown           0      800    X.X.X.X    Y.Y.Y.Y    
> 0         0         ip          0      0           1           58
> 0           0           unknown           0      800    X.X.X.X    Y.Y.Y.Y    
>  0         0         ip          0      0           1           58
> 
> The above flows are mpls double tagged. I did a binary dump on the NF packets 
> and I made sure that the data is actually sent by the router.
> 
> The question is this: is there a simple way to make it recognize the fields 
> that I want? Then, when it does, can it apply aggregation and accounting on 
> them as well?
> _______________________________________________
> pmacct-discussion mailing list
> http://www.pmacct.net/#mailinglists

_______________________________________________
pmacct-discussion mailing list
http://www.pmacct.net/#mailinglists

Reply via email to