Partition.stop: When called, will remove supervisord configuration and update supervisord.
Only started partition has supervisord configuration files.

slapgrid: If `getRequiredComputerPartitionList` fail because or network issue,
buildout cannot process instance. Slapgrid will start any partition that has supervisord configuration
file and exit. This feature is required for `slapos node boot` when computer is offline.

cli.format: parameter `--ignore_network_errors` can be set to ignore exception while
posting xml file to master.

cli.boot: Now call `slapos node format` with `--ignore_network_errors`, then remove
partition timestamp and let cron launch `slapos node instance`. Partition will be
started if there is network issue.

manager.prerm: because of the new behavior of Partition.stop, the rm script
will be killed each time `slapos node report` runs. So the prerm script is
generated in a custom supervisord configuration. Samething for `manager.portredir`.