[j-nsp] Inline IPFIX sampling rate

Euan Galloway euan+j-nsp at galloway.cc
Tue Jul 25 13:39:28 EDT 2017


Let's try that again, but to the list this time...

*sigh*

Euan

On 25 July 2017 at 10:30, Pavel Lunin <plunin at gmail.com> wrote:

> AFAIR sampling rate config hadn't had any meaning for inline IPFIX until
> some version of JUNOS. It used to be always 1, no matter what you have in
> the config.
> Question 1: If correct, which version it was when it's changed?
>

Although often said/documented/quoted, I've never actually seen this
behavior (rate = 1 regardless of configuration).
I'm *fairly* (but not 100%) sure that this didn't happen in late 11.x (even
though referenced in MX book v1)
It *certainly* doesn't happen in 12.3+


> As of the 2nd edition of the MX book, "input rate" must follow the formula
> “input_rate=2^n-1”. But it's somewhat poorly explained what it means.
> Question 2: Do I correctly understand that the value, configured with
> ”sampling instance <blah> rate <x>”, must conform to this rule. Here <x> is
> not the index of power of two (aka n in the formula above), it is the
> actual sampling rate (result of the formula), right?
>

Right below the formula is a table of valid rates (described as
"input-rate" throughout, rather than "rate")
Shown as valid are 1 / 3 / 255 / 1023,

>
> Question 3. Given that I correctly understand the previous point, what if I
> configured something that does not conform to this rule e. g. "rate 100"? I
> suspect that it should lead to "rate 1" being programmed in the PFE but
>

It works (tested on 15.1F), no obvious difference in the % error for
setting rate to 2 / 3 / 4 / 5 / 7 / 10.
(Even though 3 and 7 are somehow "valid" and the rest not).

I suspect in some prior version it rounded up to the next valid number
(since that's what commonly happens for other similar "you asked me to do
something I can't, so I'll try my best to get close"). Also, I'm sure I've
seen that behavior described by other users.


> can't find a way to check this out. Anyone knows a uKernel command to
> verify what is actually applied at the PFE level? It seems that "show
> sample instance summary" just replicates the config values.
>
Since it IS actually using the configured value, it's possible this IS the
value programmed into the hardware.
(but no, I don't know another command).
I'll see what it shows on other versions, perhaps it used to round (I
really doubt it ever just ignored you and did "1" if you missed the magic
number).


More information about the juniper-nsp mailing list