I have two Windows slaves. When I run pipeline jobs on these slaves I get tons of error messages

Cannot contact Windows7: java.io.IOException: Remote call on Channel to /172.17.xxx.xxx failed

Yet the job finishes successfully. A minimal example to reproduce is

node('windows') {
    stage('Test') {
        sh '''
            for ((i=0; i < 100; i++)); do
                echo $i
                sleep 1
            done
        '''
    }
}

Executing the same shell script via a freestyle job on the same Windows slave works without any error messages therefore it's very likely that this is a bug in the pipeline implementation.
Restarting slaves and master doesn't change anything. I also couldn't find any more information in log files (or I am looking at the wrong place)