Re: [Bloat] Router congestion, slow ping/ack times with kernel 5.4.60

[Sebastian Moeller <moeller0@gmx.de>]

Hi Thomas,

Probably not very helpful, but iputils' ping has the -D option that is described as:
      -D
           Print timestamp (unix time + microseconds as in gettimeofday) before each line.

this potentially allows to be translated into kernel time to pinpoint events in time? Not that I would know how to translate times here and if perf and friends actually store timestamps....

Best Regards
	Sebastian


> On Nov 7, 2020, at 18:01, Thomas Rosenstein via Bloat <bloat@lists.bufferbloat.net> wrote:
> 
> 
> 
> On 7 Nov 2020, at 17:46, Jesper Dangaard Brouer wrote:
> 
>> 
>> Did the latency issue happen during this this perf recording?
>> (it is obviously important that you record during the issue)
> 
> yes it was recorded during the transfer, while it exhibited the issue
> 
> I let the transfer run some time, then did the perf extract, stopped the transfer afterwards.
> 
>> 
>> Finding a single latency event/issue in perf.data is like looking in
>> for needle in a haystack.  It does give hints and e.g. tell me details
>> like what part of the kernel are used/activated.
>> 
>> I can see that you have likely configured XPS (Transmit Packet
>> Steering). Is this on purpose?
> 
> Not on purpose, I assume it comes from the Mellanox Card.
> 
>> 
>> -- 
>> Best regards,
>>  Jesper Dangaard Brouer
>>  MSc.CS, Principal Kernel Engineer at Red Hat
>>  LinkedIn: http://www.linkedin.com/in/brouer
> _______________________________________________
> Bloat mailing list
> Bloat@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/bloat