Closed
Description
Bug description
The ws-manager contains 100 workers to handle the events from the workspace pod and volume snapshot.
If there are over 100 VolumeSnapshots are ready to the ws-manager be restart, the ws-manager can't handle any pod and volume snapshot event because all the workers hang forever.
Steps to reproduce
- Create over 100 VolumeSnashots.
- Try to open the workspace, the workspace pod can't be up and with
Error
state, and the workspace pod log is{"@type":"type.googleapis.com/google.devtools.clouderrorreporting.v1beta1.ReportedErrorEvent","error":"socket did not appear before context was canceled: errors of connect to ws-daemon","level":"error","message":"cannot connect to daemon from ring0","ring":0,"serviceContext":{"service":"workspacekit","version":"commit-8d009263e19017f630e6f12436d3aa9186b8355c"},"severity":"ERROR","time":"2022-09-13T06:33:39Z","workspaceId":"gitpodio-empty-egk6yo5d3ix"} {"level":"info","message":"ring0 stopped","ring":0,"serviceContext":{"service":"workspacekit","version":"commit-8d009263e19017f630e6f12436d3aa9186b8355c"},"severity":"INFO","time":"2022-09-13T06:33:39Z","workspaceId":"gitpodio-empty-egk6yo5d3ix"}
Workspace affected
No response
Expected behavior
The workspace can be started.
Example repository
No response
Anything else?
https://gitpod.slack.com/archives/C04245JPHKL/p1663223540516549