devspace sync not working with remote K3S Cluster
What happened?
I want to use devspace sync to copy some files to a K3S Cluster on a remote Root-Server in a CICD-Pipeline.
The copy process never works and close with different errors like:
E0102 15:15:50.925683 150226 v2.go:104] use of closed network connection
Or:
E0102 15:13:49.798803 149064 v2.go:104] io: read/write on closed pipe
Or:
E0102 15:12:53.564798 148463 v2.go:104] write tcp 192.168.178.23:55576->89.58.43.43:6443: use of closed network connection
I can copy the same directory to a local minikube cluster running on my laptop. Are there any requirements for the cluster or Root-Server to make the file transfer work? Can I get more information why the transfer failed?
I just use the devspace sync command without a devspace.yaml
Local Environment:
- DevSpace Version: [use
devspace --version] -devspace version 6.3.14 - Operating System: windows | linux | mac -
linux - ARCH of the OS: AMD64 | ARM64 | i386 -
AMD64Kubernetes Cluster: - Cloud Provider: google | aws | azure | other -
Netcup Root Server with K3S - Kubernetes Version: [use
kubectl version] -v1.30.4+k3s1
I nailed the Problem down to the image running inside the Cluster.
I used the tomcat:9-jdk21-corretto image with amazonlinux Base Image.
By switching to tomcat:9-jre21-temurin image with Ubuntu Base, it is working now.
Are there any requirements regarding the Container Image which are missing for making the sync work with amazonlinux?
@mwagner5310
It sounds like the container might be restarting before the sync can complete. devspace sync won't override the container entry point like devspace dev typically does IIRC. If the container entry point depends on any of the synced files, that could be a cause.
The Devspace project currently lacks enough contributors to adequately respond to all issues. After 90d of inactivity, the issue is closed. You can re-open this issue if you still want help.