Describe the bug
First error in logs are "ERRO Error while accepting connection on 0.0.0.0:3000: Os { code: 24, kind: Other, message: "Too many open files" }, task: network", after a while its starts spamming "ERRO cannot send PullHeaders request to network: send failed because receiver is gone, task: block" and node stops receving blocks
Mandatory Information
To Reproduce
Steps to reproduce the behavior:
Dont know how to reproduce, happens after the node has run a few hours
Expected behavior
No error
Additional context
See appendix for output logs and strace logs. I have four examples as I happens every time I try to run my node.
Thank you for the captures, will investigate.
Massively duplicate connections are still present; working on a better fix.
I'm running 0.6.0 with strace, will serve you fresh logs when(/if?) it crashes :)
Error just happened on my other server, running a other setup with a old-style mechanical HDD, dont know if it is relevant. Running v0.6.1
Here are the logs:
err_24.zip
Just happened on my other server, so its not hardware related as this one uses a SSD. Still on v0.6.1
Logs:
err_24_2.zip
@bjarnekvae it's not storage related, it is a network subsystem issue as indicated in the log : task: network"
One probable cause has been fixed in #981 (and fixes of the fix that have followed).
Please retry with version 0.6.5 or later.
running 0.6.5 with strace now
This error just came back in v0.8.0-rc2
Strace log:
tooManyFiles.zip
@mzabaluev This is still relevant for v0.8.0-rc5
ERRO Error while accepting connection on 0.0.0.0:3001: Os { code: 24, kind: Other, message: "Too many open files" }, task: network
version: jormungandr 0.8.5-3db06807
Most helpful comment
Massively duplicate connections are still present; working on a better fix.