Yesterday the probe disconnected and i have not been able to get it back online. I have tried rebooting the probe, replacing the USB power adapter and micro USB cable. The device will power up, get blinky lights on the network port and a flashing RED led on the “front” of the probe.
it also does appear to still obtain a DHCP ipv4 address on my LAN, which i am able to ping (192.168.1.X/24 LAN network).
there have been no other network changes. I run the probe behind a PFSENSE box on my network – there were no changes to my IPv4/IPv6 configuration, NAT settings or ISP connectivity.
Please let me know if there is any other troubleshooting that I can perform or if the probe needs to be replaced. As i write this post, I have the probe powered off (I’m going to leave it powered off for ~10 minutes) then i will power it back up.
if these V5 hardware probes are no longer recommended to use, I will switch to a software driven probe on one of my linux servers.
Your probe was (un)lucky enough to be randomly chosen to connect to a new controller, which had a compatibility issue. It has now connected to another endpoint and should behave as normal.
In general we try to test new infrastructure using test probes, but sometimes there are issues that only appear in the “real world”. Apologies for the downtime!
I am experiencing a similar issue with my RIPE Atlas probe (Probe ID: 65180). My ISP went down around 1 AM on 11 July and service was restored at 11 AM the same day. Since then, the probe has remained in a disconnected state.
I have another probe in the same network that is functioning normally, so I have ruled out internal network issues. The affected probe is receiving power, and the Ethernet lights are active (green constant, amber blinking). The SOS history reflects each instance I disconnect and reconnect the probe. Additionally, the probe has been tagged with “Firewall problem suspected” and “Trying to connect.”
After an DSL reconnect my probe is also not able to connect. Still marked as offline. I tried IP reconnects, DSL reconnects, disconnected the probe for 1h…
Same here (v5 [probe, 61817). After a disconnect for a few minutes on the network side the probe is no longer connecting.
I see v4 and v6 communications with reg02.atlas.ripe.net and atlas-ui.ripe.net, but no connections to an actual controller and the status is red for the past ~2 hours. Other machines work fine for both v4 and v6 traffic.
@camin
Yeah something is not right on the RIPE Atlas infrastructure it seems, 5 days in a row I’ve now problems keeping the probe (v4, 50918) staying online, it’s online for ~10 minutes then marked as offline.
Connection traffic in/out is still visible on the probes page.
The problems started 2024-07-10
The status page says “likely firewall issues” but I see successful connections with ssh traffic to the registration servers. I can manually connect to the recent controllers the probe was connected to, so firewall should really not be the issue.
Probe is now offline for the past 20h with no change I know of in the network here.
Installed a software probe (1008577) as that is easier to debug. Came online without any issues from the same network. So this must be something wrong with the hardware probe.
Yeah, seeing the exact same behavior here (SOS history updates, probe functions are executed, but no connection to controller). Also with another probe in the same network working.
Reporting for the dev team: we are making changes to the infrastructure that deals with probes, and in some cases this can cause (temporary) connection issues. Many of the probes reported here were caught in that situation. We apologise for these inconveniences.
It took a while but after the restart it has recovered and has communications again, thank you.
Hello @robertk my probe is not working, with the LED flashing red. I have restarted it, but it remains in the same state.
Disconnected for 7 hours. Are the changes still going on? If this is correct, what information should I provide to verify that the probe I have is affected by these changes? Thanks!
Regards!
Changes are still ongoing to my knowledge. I’m picking this up because Robert is enjoying his holidays
Your probe indeed looks offline. Can you give me the exact flash pattern? (time on, time off).
First thing to check is whether the power supply and cable are still ok. It depends a little on the blink pattern if that is relevant in this case. Next, make sure the unit gets an IP address (verify the MAC address in the DHCP leases). Check your internet connection for good measure. Any changes there?
It took a while but after the restart it has recovered and has communications again, thank you.
Hi @mstam , it’s fixed after the reboot, I should have mentioned it in a new post.
Thank you very much for the data for the tests, I’ll save them to perform tests in case I need them in the future, thank you very much for all the help!
I also have a v5 probe (#60623) that has been erratic. It randomly says it’s disconnected, even though it’s still able to send those SOS messages and my actual internet is still working.
Often rebooting the probe and waiting a bit makes it come back online, but then it disconnects after some random period of time.
Would this be a problem with my probe, or is it related to the infra changes mentioned here?
FWIW, I also have a v4 probe attached to another ISP that’s facing no issues.
Exactly same problem on my Probe 54080, successufully connected to the LAN and the NET, visible on my router, pingable from my router.
But since May 15 (more than 85 days), it fails to connect to the RIPE servers which reports it as “disconnected”. The probe has always been online, and I have rebooted it multiple times (plugging it off and on after a few seconds to reboot it), and nothing changes.
I was never notified of the failure. The probe seems to work and has its two external LEDs (on the Ethernet port) flashing slowly in green (1 blip every few seconds), but the internal LED (visible through the air vents) is constantly red.
I’ve sent a mail to the support, including snapshots of the status page on RIPE, as well as a snapshot of my fiber router, showing that there has been NO network change at all since over one year (same IPv4, same IPv6), and my fiber connection is working flawlessly).
You’ve visibly incorrectly reconfigured the probe on May 15 and rebooted it, and it worked only for 3 hours and then stopped working, which cannot connect on your specified server.
Powering the probe off and oon (on different days or hours) does not change anything after many attempts…
Apparently the RIPE server configured by you in your probe is rejecting the connection and since more than 2 months, this has never been fixed. Do you monitor when probe access are rejected? Why don’t you look at probes that should be online, but that your server rejects? May be it’s one of your server that is no longer working since months and you forgot to fix the probe settings)!
