You are here: Home > Participate > Join a Discussion > RIPE Forum
RIPE Forum v1.4.1

RIPE Atlas

Threaded
Collapse

[atlas] One of my hosted probes don't accept measurements

User Image

Evgeniy S

2020-05-06 19:50:52 CET

Dear RIPE guys,

one of probes hosted by us https://atlas.ripe.net/probes/35439/ shows as
connected and good to go, but when I create any measurements with this ID I
see "1 Requested / None yet Actually Participating"

Please advice what's wrong with it and how to troubleshoot?

--
With regards,
Eugene Sudyr
User Image

Evgeniy S

2020-05-07 07:27:07 CET

Later it says "NO SUITABLE PROBES from
2020-05-06T17:31:28Z
 to
2020-05-06T17:40:01Z"

On Wed, May 6, 2020 at 8:50 PM Evgeniy S. <eject.in.ua _at_ gmail _dot_ com> wrote:

> Dear RIPE guys,
>
> one of probes hosted by us https://atlas.ripe.net/probes/35439/ shows as
> connected and good to go, but when I create any measurements with this ID I
> see "1 Requested / None yet Actually Participating"
>
> Please advice what's wrong with it and how to troubleshoot?
>
> --
> With regards,
> Eugene Sudyr
>


-- 
--
With regards,
Eugene Sudyr
User Image

Viktor Naoumov

2020-05-07 09:09:57 CET

RIPE NCC staff member

Hi Eugene,

I've checked the activity log of your probe and it makes me strongly 
suspect that the USB flash is corrupt, either the filesystem of the 
drive itself.

Please follow these steps to recover your probe: 
https://labs.ripe.net/Members/philip_homburg/troubleshooting-ripe-atlas-probes-usb-sticks

It can happen sometimes that a probe is shown green but it refuses to 
accept commands.


Best regards

RIPE NCC R&D

/vty




On 5/6/20 7:50 PM, Evgeniy S. wrote:
> Dear RIPE guys,
>
> one of probes hosted by us https://atlas.ripe.net/probes/35439/ shows 
> as connected and good to go, but when I create any measurements with 
> this ID I see "1 Requested / None yet Actually Participating"
>
> Please advice what's wrong with it and how to troubleshoot?
>
> --
> With regards,
> Eugene Sudyr


User Image

Evgeniy S

2020-05-07 09:28:51 CET

Thanks Viktor,

It was also my understanding of root cause of this problem.

I did test again selecting 500 probes in EU region: 500 Requested / 429
Actually Participating

Does it mean that all of them have same issue as mine of there are other
reasons why selected probes are not participating?

Do you know if there is any plan to make any self-check on probe to
diagnose this kind of problems proactively?

--
Evgeniy

On Thu, May 7, 2020 at 10:10 AM Viktor Naumov <vnaumov _at_ ripe _dot_ net> wrote:

> Hi Eugene,
>
> I've checked the activity log of your probe and it makes me strongly
> suspect that the USB flash is corrupt, either the filesystem of the drive
> itself.
>
> Please follow these steps to recover your probe:
> https://labs.ripe.net/Members/philip_homburg/troubleshooting-ripe-atlas-probes-usb-sticks
>
> It can happen sometimes that a probe is shown green but it refuses to
> accept commands.
>
>
> Best regards
>
> RIPE NCC R&D
>
> /vty
>
>
>
> 
> On 5/6/20 7:50 PM, Evgeniy S. wrote:
>
> Dear RIPE guys,
>
> one of probes hosted by us https://atlas.ripe.net/probes/35439/ shows as
> connected and good to go, but when I create any measurements with this ID I
> see "1 Requested / None yet Actually Participating"
>
> Please advice what's wrong with it and how to troubleshoot?
>
> --
> With regards,
> Eugene Sudyr
>
>
>

-- 
--
With regards,
Eugene Sudyr
User Image

Viktor Naoumov

2020-05-07 11:41:44 CET

RIPE NCC staff member

Eugene,

There is never a guarantee that you will get all the probes requested, 
since some of them may be faulty, disconnecting, probe busyness 
limiting, etc. In such cases you should request a bit more to make the 
number of participants be closer to your desired quantity.

The scheduling process is a two-stage. First stage happens on 
nearly-real-time probe data on the central scheduler then it is split 
and passed down to controllers, where the probes connected to and that 
have the real picture of the probes state. Based on that knowledge 
controllers filter the probe list further.

There is the probe diagnostic system that use some heuristics to detect 
faulty probes and tag them, it works quite well but again it is very 
difficult to cover all the cases. We are working on improving that.

Best regards
/vty

On 5/7/20 9:28 AM, Evgeniy S. wrote:
> Thanks Viktor,
>
> It was also my understanding of root cause of this problem.
>
> I did test again selecting 500 probes in EU region: 500 Requested / 
> 429 Actually Participating
>
> Does it mean that all of them have same issue as mine of there are 
> other reasons why selected probes are not participating?
>
> Do you know if there is any plan to make any self-check on probe to 
> diagnose this kind of problems proactively?
>
> --
> Evgeniy
>
> On Thu, May 7, 2020 at 10:10 AM Viktor Naumov <vnaumov _at_ ripe _dot_ net 
> vnaumov _at_ ripe _dot_ net>> wrote:
>
>     Hi Eugene,
>
>     I've checked the activity log of your probe and it makes me
>     strongly suspect that the USB flash is corrupt, either the
>     filesystem of the drive itself.
>
>     Please follow these steps to recover your probe:
>     https://labs.ripe.net/Members/philip_homburg/troubleshooting-ripe-atlas-probes-usb-sticks
>
>     It can happen sometimes that a probe is shown green but it refuses
>     to accept commands.
>
>
>     Best regards
>
>     RIPE NCC R&D
>
>     /vty
>
>
>     On 5/6/20 7:50 PM, Evgeniy S. wrote:
>>     Dear RIPE guys,
>>
>>     one of probes hosted by us https://atlas.ripe.net/probes/35439/
>>     shows as connected and good to go, but when I create any
>>     measurements with this ID I see "1 Requested / None yet Actually
>>     Participating"
>>
>>     Please advice what's wrong with it and how to troubleshoot?
>>
>>     --
>>     With regards,
>>     Eugene Sudyr
>
>
>
>
> -- 
> --
> With regards,
> Eugene Sudyr