If you know a disaster is about to happen, you’re a step ahead in dealing with it. Or simply preventing it. After all, who ever said you had to wait for a disaster to happen to practice good DR?
Today’s IT systems often include mechanisms for signalling if things start to go awry. Error rates in hardware can be monitored and analysed automatically to flag situations that are out of bounds.
IT staff can then plan downtime to replace suspect components, switching to backup systems to pick up the load while repairs are being made. But what if the cry for help was noticeable without the help of data analytics?
In the case of hard disk failure, warnings signs may be heard or felt before failure, even if data is still being stored and retrieved without error.
Loud humming or grating sounds and excessive vibration are indications that something is amiss. This is one of the few advantages that hard disks retain over solid state disks, but you need to be at hand and pay attention to what you hear or feel.
On a bigger scale, rising or uneven temperatures in a data centre can herald problems. This may be either because the air conditioning system is faulty, or because systems are improperly placed for balanced cooling (often after they have been moved around, changing the original layout).
If you’re using cloud computing services, you won’t have the luxury of ambling down each aisle in the computer centre. Remote monitoring and diagnostics will be your only recourse, and those cries for help will all be virtual.
However, so typically will your machines (virtual machines) and as SSD continues to make inroads into global storage capacity, virtual voices will be the only ones left anyway. The need to listen for them and to them will not change, however, simply the way the listening is done.