Fix kubeturbo panic when it is stopped on command line

[ading1977]

Problem

kubeturbo panic when stopped on command line:

I0624 11:57:56.293012   77936 tap_service.go:111] Begin to stop TAP service.
I0624 11:57:56.293022   77936 mediation_container.go:86] [CloseMediationContainer] Closing mediation container .....
I0624 11:57:56.293031   77936 tap_service.go:113] TAP service is stopped.
I0624 11:57:56.293041   77936 kubeturbo_builder.go:447] Kubeturbo service is stopped.
panic: send on closed channel
goroutine 346 [running]:
github.com/turbonomic/turbo-go-sdk/pkg/service.(*TAPService).getJwtToken(0xc000a937a0, 0xc000584f78, 0xc000584f68)
        /Users/mengding/go/src/github.com/turbonomic/kubeturbo/vendor/github.com/turbonomic/turbo-go-sdk/pkg/service/tap_service.go:60 +0x56
created by github.com/turbonomic/turbo-go-sdk/pkg/service.(*TAPService).ConnectToTurbo
        /Users/mengding/go/src/github.com/turbonomic/kubeturbo/vendor/github.com/turbonomic/turbo-go-sdk/pkg/service/tap_service.go:106 +0x265
Process finished with the exit code 2

Cause

This is caused by a race condition, where kubeturbo may write to a closed channel (i.e., jwTokenChannel) during clean up.

Solution

• Break out of the loop of getJwtToken when detecting the close of refreshTokenChannel channel. This way, there is no chance for kubeturbo to write to jwTokenChannel when refreshTokenChannel is closed.

[coveralls]

Coverage decreased (-0.03%) to 48.258% when pulling 2d6a43a on ading1977:bugfix/OM-86907 into 5e8ddac on turbonomic:master.