My FR24 feeds keeps dropping every few days while others run fine!

[xpro]

Hi,


I would like to ask for some help on troubleshooting the issue I have. Every few days I have outages on my fr24 feed..

Im running a dump1090 with multiple feeds, flightaware, planefinder, adsb exchange and most importantly FR24 feed.

The problem is that while I get outages on the fr24 feed, the other providers are working totally fine without disruption.

In this case I have to either reset the Pi or after about 5hr it will come back to life again..

Can anyone suggest what the problem could be here?

Thanks

[recapa]

Hi,

I can confirm same behaviour with my 2 feeders in last couple of days. Irregular they stop to be accessible and restart of the fr24feeder is necessary to get service running again. My 2 feeders are running in two different locations and on different networks. They have been feeding since months without any issue until short time ago.

I can confirm all other feeding application running on same Raspberries continue to be accessible and feed the data without any issue.

As the client 1.0.14-11 has not been updated since while this has to be some change on FR24 server side which causes freeze of client by communication.

Regards, Jan

[Oblivian]

1.0.14-11 ? That's from 2015 btw. Not just a while. Up to 26-5 now.

[recapa]

My mistake by copy&paste, installed version on both PIs having issues with communication to FR is 1.0.26-5

From timing perspective the issue occurrence shows it had to be introduced with auto update to 1.0.26-4 (2020-09-01) or 1.0.26-5 (2020-09-15).

I have decided to disable IPv6 on one of the feeders and will see if this will fix the problem and keep issues on other one.

Thanks, Jan

[Oblivian]

Not much help I'm afraid.

Each time I try to suggest info is collected /investigated locally to rule something common unfortunately I get accused of blaming users or being a denier that theres a server issue

Need to collate what the commonality is between everyone seeing it to find a path of likely cause. Location (effects server node it attaches to) Pi version (incase the new ones are causing a segfault) and enable the log for a while and review when it cuts out

I force upgraded my version (I have auto update off) from 25- something to the latest one to see what everyone was complaining about and if I could replicate it. And even the outtage I did to take it down and upgrade barely showed up as a downtime.

At no point since have I had to restart or had an outtage.

Assuming everyone is looking at the stats page rather than the output of the feeder status?

fr24feedOK.JPG


Really need some logging enabled and checked when it happens to see if console is reporting anything or not. Lots of variables.

here for instance - feed.fr24.com (one of the locations I think the feeder connects to) comes back as 104.19.208.130 - this will be different per region as it is cloudflare

On startup you can see the others it connects to

Direct to login /data server at FR - 185.218.24.22:8099

Configured ReceiverACSender: 185.218.24.22:8099,185.218.24.23:8099,185.218.24.2 4:8099,

and I think the MLAT server - udp://australia.fr24.com:19788 - Amazon Australia, again, this will vary by region

But you see, I'm also not a USB stick or dump1090 user...

[recapa]

Actually stat page is something secondary for me.

As xpro wrote in first post, the fr24 become unresponsible and only restart fix the issue. While all other non FR feeders keep working without issue and they are taking data from same port on same PI. I continue to look for the issue, but I don't have answer so I have been looking for people with same issues in last days and hoped to dfind some direction and I'm obviously not alone.

I will see if i can find something to give me better answer.

[Oblivian]

It depends on who you ask and how much info is given.

If It was global their social media blows up and lots report here. But so far i only count about 4. All with slightly diff methods of testing and scenarios.

Another user has reported it, but as of today reports its different to what you describe

It started off with 'fr24feed stopping'

But his local traffic is now found it does go offline, and even fa reports blocked local ports.

So although they sound the same. When you dig deeper it may not always be the case

In your case you need to focus on the logs from fr24feed if others confirmed. enable them in settings and restart it for a while. Recheck them at time of failure

cat /var/log/fr24feed/fr24feed.log

It may be their ipv6 changes need to be better configured, or disabled so it resolves right.

But it will also show you the ips it connects to if they differ to mine above

[Oblivian]

I just noticed you edited after my reply about testing ipv6. That was my only conclusion too as it was the most recent version changes.

[xpro]

For me the same issue. I cannot understand this outages .

And same as before, my other feeds are working without any issue. That is for the periods of fr24 outage, Flightaware is still running fine and so is the planefinder..


Edit: Quick google search reveals that this is currently an ongoing issue with fr24 servers.

They only pushed a new update and it seems maybe they are trying to fix something ,

Lost of posts on flightaware too about their outages..


Screenshot 2020-09-29 at 19.44.07.png

[Oblivian]

For me the same issue. I cannot understand this outages .

Easy

Send an update to 'fix' ipv6 and dns lookup issues on everyones feeder - makes it worse and breaks if auto update enabled.

At the samw time add to that half of russia/china/whichever countries computers probably sending 10mb ping packets to every server fr24 IP known, and you have a DDoS that is happening

https://www.cloudflare.com/learning/...a-ddos-attack/

BUT

Again, this is a different issue to what people describe/give data for.

That only effects the server receiving stuff. Not someones local app stopping. Still need to check for common reasons for that. Despite getting an offline message now and then due server error, my feeds been fine. Which isn't what others are seeing

[Oblivian]

There's some very distinct scenarios that seem to be appearing that need to be cleared up in detail. Everyone is mashing their experiences into threads of others that likely only partly share common trend and causing confusion

1
updated feeder blocked web access
- a new one was released

2
Feeder server stops responding - ddos, gaps in feed STATS only.
also needs local traffic received status verification

3
feeder process stops - can be the update, or local configuration
Needs to be confirmed by logs what/when
Needs to compare if dvbt or particular config
need more to test old version
Depends on ddos ending