What action must be taken if undesirable results occur after deploying a patch to virtual machines?

When undesirable results occur after deploying a patch to virtual machines, the appropriate action to take is a rollback. Rolling back a patch involves reverting the virtual machines to their previous state before the patch was applied. This is crucial in maintaining system stability and ensuring that operations can continue efficiently without the issues introduced by the patch.

Rollback is particularly important in cloud environments where uptime and reliability are critical. It allows administrators to quickly mitigate any problems that arise from the patch, thus minimizing downtime and the impact on users. By restoring the previous version of the software, it also helps to identify whether the patch itself was the cause of the issues, ensuring that any necessary troubleshooting can be performed effectively.

Other options such as "update," "reinstall," and "restore" may not directly address the immediate need to revert to a stable state. An update implies applying a new version or patch rather than reverting to a previous one. Reinstalling the software could lead to more downtime and potentially more complications, while restoring typically refers to restoring a state from backups, which may not necessarily be as efficient as a rollback in this context.