Easy to use system to shut down your infrastructure, when loosing the power. After an initial setup, newly added virtual servers are automatically recognized and integrated into the program. Optionally, the order of the servers can also be changed so that the most important or sensitive systems, such as databases and mail systems, are shut down first. Systems that should be running until the end will then shut down last. The shutdown of virtual systems is released with a configurable time difference in order to avoid an overload of your storage system. The entire system can be tested in a debug mode without having an actual shutdown. You will receive an email regarding any communication problems with hosts or UPS systems, also power losses and power recoveries. You can therefore be reset assured that this system will do exactly what is expected in the case of emergency.
A power failure can always cause major problems for server systems. This is especially true for virtual infrastructures, since a large number of virtual machines are affected at the same time. Server systems switched off without a proper shutdown, whether real or virtual, have good chances of taking intensive installation damage or even worse, malfunctioning files and data loss. As an example, databases or mail systems may then become unusable after an outage. Because of this, continual power supplies (UPS) are widely used as the standard in server environments. These power supply systems are capable of filtering short instabilities only lasting a few seconds without affecting your server systems. During longer outages of minutes or even hours however, servers need to be shut down in a controlled manner. The attempt to achieve this controlled shut down is normally done using additional programs or scripts which are usually bundled with UPS hardware.
These scripts unfortunately have several shortcomings: Firstly, they must be adapted during the reconfiguration after an environment. If changes to your virtual infrastructure have not been properly embedded into these scripts, nothing will happen once the power source dies. These changes also need to be tested in advance. In addition to this, using these bundled solutions shut down all virtual servers without any consideration of the succession in which the shutdown should proceed. This can either significantly lengthen the general shutdown process or lead to inconsistencies with in data files. The shutdown of an Exchange server, for example, takes about three times longer if the domain controllers have already shut down. A simultaneous shutdown of many virtual servers in the same storage unit will also cause a system performance overload. Due to this, the shutdown process will be much longer – in some cases longer than the battery in your UPS is able to support.
