Panic when consuming a (big) notification 2.0 backlog

[ButKor]

I've consumed a Notification 2.0 backlog of ~ 150k messages via: c8y notification2 subscriptions subscribe --name {my subscription name} --subscriber {subscriber name}.

After ~ 100k messages this panic got thrown:

panic: concurrent write to websocket connection
 
goroutine 1 [running]:
github.com/gorilla/websocket.(*messageWriter).flushFrame(0xc0007845a0, 0x1, {0x0?, 0x0?, 0x0?})
	github.com/gorilla/[email protected]/conn.go:632 +0x4b8
github.com/gorilla/websocket.(*messageWriter).Close(0xa3d7d8?)
	github.com/gorilla/[email protected]/conn.go:746 +0x35
github.com/gorilla/websocket.(*Conn).beginMessage(0xc000737ce0, 0xc00081e840, 0x1)
	github.com/gorilla/[email protected]/conn.go:493 +0x47
github.com/gorilla/websocket.(*Conn).NextWriter(0xc000737ce0, 0x1)
	github.com/gorilla/[email protected]/conn.go:535 +0x3f
github.com/gorilla/websocket.(*Conn).WriteMessage(0xc18e9a9c60b784d0?, 0x29f87f4050?, {0xc0002bc000, 0x10, 0x1000})
	github.com/gorilla/[email protected]/conn.go:788 +0x137
github.com/reubenmiller/go-c8y/pkg/c8y/notification2.(*Notification2Client).SendMessageAck(0xc000134b00, {0xc0002bc000, 0x10, 0x1000})
	github.com/reubenmiller/[email protected]/pkg/c8y/notification2/notification2.go:388 +0x113
github.com/reubenmiller/go-c8y-cli/v2/pkg/cmd/notification2/subscriptions/subscribe.(*SubscribeCmd).RunE(0xc0002e1cb0, 0xc0004cbb08, {0x0?, 0x0?, 0x0?})
	github.com/reubenmiller/go-c8y-cli/v2/pkg/cmd/notification2/subscriptions/subscribe/subscribe.manual.go:175 +0x6fc
github.com/reubenmiller/go-c8y-cli/v2/pkg/cmd/notification2/subscriptions/subscribe.NewSubscribeCmd.NewSubCommand.func4(0xc0004cbb08, {0xc0005bfb00, 0x0, 0x4})
	gith:55 +0x62
github.com/spf13/cobra.(*Command).execute(0xc0004cbb08, {0xc0005bfac0, 0x4, 0x4})
	github.com/spf13/[email protected]/command.go:983 +0xaca
github.com/spf13/cobra.(*Command).ExecuteC(0xc0002e8308)
	github.com/spf13/[email protected]/command.go:1115 +0x3ff
github.com/spf13/cobra.(*Command).Execute(...)
	github.com/spf13/[email protected]/command.go:1039
github.com/reubenmiller/go-c8y-cli/v2/pkg/cmd.MainRun()
	github.com/reubenmiller/go-c8y-cli/v2/pkg/cmd/cmd.go:124 +0x1d0
main.main()
	./main.go:8 +0xf

When restarting the subscription, the rest of the backlog was consumed fine. Not sure if there is something to "fix", but I guess the panic should be handled to e.g. show a proper error message and maybe even retry after some time?

[reubenmiller]

Fixed in v2.41.1