syslog size for public peer
d47081 opened this issue · comments
hello there!
I found that my public peer server (yggdrasil 0.5) starting to generate massive /var/log/syslog files with following content
Nov 19 00:01:23 yggdrasil[387592]: 2023/11/19 00:01:23 Connected inbound: 201:23b4:991a:634d:8359:4521:5576:15b7@94.140.114.241:39284, source 94.140.114.241:4708
Nov 19 00:01:23 yggdrasil[387592]: 2023/11/19 00:01:23 Disconnected inbound: 201:23b4:991a:634d:8359:4521:5576:15b7@94.140.114.241:39284, source 94.140.114.241:4708; error: ErrBadKey: Expected KEY, Found KEY
...
is there any way how to reduce it with debug level or that's bad idea?
I'm using default deb package with systemd without any changes in auto-generated config
Just size of logs grow up to 1 Gb, I think, that's issue for VPS
You could switch to -loglevel warn, which will silence most of the logging, but this sounds like you probably need to look at why your system is letting your syslog grow to that extent. rsyslog etc have options for this.
logrotate package should be able to handle this.
Thanks for advice, it's maybe happens when node become closer to the network root so just for info
note: ygg 0.5 got rid of the DHT spanning tree so the concept of a root node is no more, it's replaced with a "core" set of routers which are the most well connected.
@neilalexander could you please explain where to add this option -loglevel warn
I have yggdrasil installed from deb package, can't find where systemd file stored or I must to recompile it with this option?
@logrotate thanks for logrotate advice, but I want to silentize non-important messages from yggdrasil just (like connection events) and keep another ones that could be important
just it happened again, my logsize grows and it causes free disk space issues..
Can anyone explain why such massive amount of disconnects happen?
My first thought was that it happens because of crashes.
But I doubt nodes can crash every 2 minutes.
(upd: Sometimes nodes reconnect every few seconds.)
So what is happening?
It is the regression (hopefully) fixed in v0.5.4. We need to wait until all public peers upgrade to 0.5.4
@d47081 You can close this issue? I hope you found where to put your option. Also people upgraded to 0.5.4
Just forgot about as moved to another server with infinitive storage ;D
Thanks for reminding!