Hi all! I'm new to RIPE Atlas, and right now using it for educational purposes :)) lately I've seen that some of my traceroute measurements do not have RTT. although it's not completely ignored by the hop (I get the data like from, size, ttl), but only rtt is missing :)) any idea how, why and when does this happen? :) for a live example, check Measurement #4458653. 48th probe's 11th hop's 1st query (counting from zero, btw) does not have rtt :) thanks in advance! side question: is there an (un)official IRC channel? :) -- Antranig Vartanian | PGP Key ID : 0xDAB81456 https://antranigv.am | https://cert.am (* do one thing and do it well *)
Hi, On 2016/07/08 11:38 , Antranig Vartanian wrote:
I'm new to RIPE Atlas, and right now using it for educational purposes :)) lately I've seen that some of my traceroute measurements do not have RTT. although it's not completely ignored by the hop (I get the data like from, size, ttl), but only rtt is missing :))
any idea how, why and when does this happen? :)
for a live example, check Measurement #4458653. 48th probe's 11th hop's 1st query (counting from zero, btw) does not have rtt :)
Maybe you can give the actual probe ID? Or better yet, a copy of the JSON result for that probe and the actual hop number in the JSON blob? Philip
Hi, sure! Probe ID: 938. Hop: 12 I uploaded that probe's json part - https://antranigv.am/misc/result_prb_id_938.json :) -- Antranig Vartanian | PGP Key ID : 0xDAB81456 https://antranigv.am | https://cert.am (* do one thing and do it well *) On 07/08/2016 01:46 PM, Philip Homburg wrote:
Hi,
On 2016/07/08 11:38 , Antranig Vartanian wrote:
I'm new to RIPE Atlas, and right now using it for educational purposes :)) lately I've seen that some of my traceroute measurements do not have RTT. although it's not completely ignored by the hop (I get the data like from, size, ttl), but only rtt is missing :))
any idea how, why and when does this happen? :)
for a live example, check Measurement #4458653. 48th probe's 11th hop's 1st query (counting from zero, btw) does not have rtt :)
Maybe you can give the actual probe ID?
Or better yet, a copy of the JSON result for that probe and the actual hop number in the JSON blob?
Philip
On 2016/07/08 12:07 , Antranig Vartanian wrote:
sure! Probe ID: 938. Hop: 12 I uploaded that probe's json part - https://antranigv.am/misc/result_prb_id_938.json :)
Hi, There is a 'late' field. Basically that means that the packet arrived after the traceroute code already timed out. The current implementation stores just one transmit time. So after it moved on, there is no way to compute the rtt. Philip
Philip, I am a bit confused, is that "late reply from the probe" or it is "late reply from router - basically timed out reply" ? Should it be treated like packet loss ? Regards. /Alex Saroyan On Fri, Jul 8, 2016 at 4:06 PM, Philip Homburg <philip.homburg@ripe.net> wrote:
On 2016/07/08 12:07 , Antranig Vartanian wrote:
sure! Probe ID: 938. Hop: 12 I uploaded that probe's json part - https://antranigv.am/misc/result_prb_id_938.json :)
Hi,
There is a 'late' field. Basically that means that the packet arrived after the traceroute code already timed out. The current implementation stores just one transmit time. So after it moved on, there is no way to compute the rtt.
Philip
On 2016/07/08 14:19 , Alex Saroyan wrote:
I am a bit confused, is that "late reply from the probe" or it is "late reply from router - basically timed out reply" ? Should it be treated like packet loss ?
It is replies received by the probe. So it is usually from routers, but it could be from the target. The reply did arrive, so technically it is not packet loss. But it makes processing a lot easier to treat them as lost. Philip
Do you mean packet was returned later then timeout value ? How long is timeout value then ? On Fri, Jul 8, 2016 at 4:24 PM, Philip Homburg <philip.homburg@ripe.net> wrote:
On 2016/07/08 14:19 , Alex Saroyan wrote:
I am a bit confused, is that "late reply from the probe" or it is "late reply from router - basically timed out reply" ? Should it be treated like packet loss ?
It is replies received by the probe. So it is usually from routers, but it could be from the target.
The reply did arrive, so technically it is not packet loss. But it makes processing a lot easier to treat them as lost.
Philip
On 2016/07/08 14:27 , Alex Saroyan wrote:
Do you mean packet was returned later then timeout value ? How long is timeout value then ?
Yes, otherwise there would be a rtt value. The per packet timeout defaults to 4 seconds but can be set explicitly when the measurement is created. Philip
Thanks On Fri, Jul 8, 2016 at 4:32 PM, Philip Homburg <philip.homburg@ripe.net> wrote:
On 2016/07/08 14:27 , Alex Saroyan wrote:
Do you mean packet was returned later then timeout value ? How long is timeout value then ?
Yes, otherwise there would be a rtt value. The per packet timeout defaults to 4 seconds but can be set explicitly when the measurement is created.
Philip
participants (3)
-
Alex Saroyan -
Antranig Vartanian -
Philip Homburg