jenkins slave startup is not robust against temporary network issues

Currently, if a Jenkins slave machine is not able to access the Jenkins server at all over boot (e.g. DNS issue or temporary network disruption), the slave script will abort and never restart until the machine is manually rebooted.

Sample error:

fetch http://ci-dev.qt-project.org//jnlpJars/slave.jar: 500 Can't connect to ci-dev.qt-project.org:80 (Bad hostname 'ci-dev.qt-project.org')