Unexpected EOF with devspace sync

[dsebastien]

What happened?

In some cases, devspace sync crashes with the following error:

[warn]   Log streaming service has been terminated
[done] √ Sync and port-forwarding services are running (Press Ctrl+C to abort services)
[error]  Fatal sync error: upstream: apply changes: apply creates: upload archive: after upload: rpc error: code = Unknown desc = untar all: untarNext: io copy tar reader: unexpected EOF. For more information check .devspace/logs/sync.log

In the sync.log file I can see the following:

{"level":"info","msg":"Upstream - Successfully processed 17 change(s)","time":"2019-11-27T00:09:50+01:00"}
{"level":"info","msg":"Upstream - Upload 50 create changes (size 1663407)","time":"2019-11-27T00:09:50+01:00"}
{"level":"info","msg":"Upstream - Upload 53 create changes (size 1669808)","time":"2019-11-27T00:09:50+01:00"}
{"level":"info","msg":"Upstream - Successfully processed 17 change(s)","time":"2019-11-27T00:09:50+01:00"}
{"level":"info","msg":"Upstream - Upload 53 create changes (size 1669808)","time":"2019-11-27T00:09:50+01:00"}
{"level":"info","msg":"Upstream - Successfully processed 50 change(s)","time":"2019-11-27T00:09:50+01:00"}
{"level":"info","msg":"Upstream - Upload 10 create changes (size 1633384)","time":"2019-11-27T00:09:50+01:00"}
{"level":"info","msg":"Sync stopped","time":"2019-11-27T00:09:50+01:00"}
{"level":"error","msg":"Sync Error on /home/sebastien/wks/project: upstream: apply changes: apply creates: upload archive: after upload: rpc error: code = Unknown desc = untar all: untarNext: io copy tar reader: unexpected EOF","time":"2019-11-27T00:09:50+01:00"}
{"level":"info","msg":"Sync stopped","time":"2019-11-27T00:09:50+01:00"}
{"level":"error","msg":"Sync Error on /home/sebastien/wks/project: upstream: apply changes: apply creates: upload archive: after upload: rpc error: code = Unknown desc = untar all: untarNext: io copy tar reader: unexpected EOF","time":"2019-11-27T00:09:50+01:00"}

What did you expect to happen instead?
Ideally no error at all (hehe), but if possible, anything but a crash of devspace. I'd like it to be more resilient so that I don't need to check the terminal where it runs to see if I need to restart it as it breaks the development cycle / concentration.

How can we reproduce the bug? (as minimally and precisely as possible)

Local Environment:

• Operating System: linux
• Deployment method: kubectl apply

Kubernetes Cluster:

• Minikube with k8s v1.16 and kubectl 1.16.2

Anything else we need to know?

/kind bug

[FabianKramm]

@dsebastien thanks for reporting this and the detailed problem description! This is a tricky issue, I think it can appear when the file locally changes while the sync is trying to pack it. I'll try to reproduce it and implement that the sync will attempt to resend in such a case.