Hi,
there seems to be NG running in the bottom… then shutdown and then it says NG requested but Legacy used because the agent doesn’t support NG. So it seems Legacy is used at the end.
Also it would seem the connection is working from the log. Would have to check with cloud guys why cloud doesn’t show it online.
Hey,
It seems like the node shows up nominally on the Cloud panel. Unless you guys patched something, it has fixed itself. Thanks for the help so far!
@eramsorgr yep your log suggested that you are indeed online. May have been some intermittent cloud issue… Or small backlog causing a bit of delay until the node state propagated trough the system
For future reference I just want to add here that I have found : this which shows the available flags for making libwebsockets.
I see a couple that may seem interesting in regards to this case if someone wants to hunt it down further:
LWS_FALLBACK_GETHOSTBYNAME
LWS_WITH_SYS_ASYNC_DNS
Especially the last one seems a bit of intriguing based on this.
I have reverted everything to stable branch with ACLK-NG so I believe that I cannot be more useful on this one but just in case someone bumps into it.
Kind regards
Hey everyone. It seems like one of our nodes is showing unreachable again! Running v1.31.0-142-nightly. Any ideas?
@eramsorgr NG or Legacy? You might have lost the ability to run ACLK-NG due to missing protobuf packages?
We are working on a bundler in the installer that will make it available also for people without protobuf.
What does netdata -W buildinfo say?
I am on Legacy.
@eramsorgr are we talking about the node that is working only with ACLK-NG previously but now it reverted to Legacy? If so reason would be a new dependency (google protobuf).
You can install protobuf protobuf-devel protobuf-compiler depending on the system (the exact package names). Also, @Austin_Hemmelgarn is working on a patch that will allow the installer to automatically bundle it for systems that do not have it (this will sort it out for people who use automatic updater as well).
Hi @underhood, no this is actually for another node that never had an issue before. It seems though the node shows up as online after a restart to the agent. No clues.
@eramsorgr could be as well some bug on the cloud side. Hard to tell. If you have an issue again and it is not characterized by error.log message (ernno 99, Cannot assign requested address) could you open a new issue.
This thread has become “any issue causing node to not connect to cloud” and I would like to prevent that.