We mostly talk tech with a technical touch. For a change, lets talk tech in ‘fairy tale’ mode for Backup, and as you all know fairy tales always begin with…
Once upon a time, there lived a family that had a ‘grandfather, a father and a son’. Their job was to remember events. The son being the youngest among the three, remembered all of the latest events on a daily basis. The father being elder to the son remembered events on a weekly basis, which finally left grandfather, the eldest, to remember events on a monthly basis. Their work seems to be pretty simple – it is ‘to remember.’ The son helped his father with his work, which led the father to help his son’s grandfather with the work of remembering.
As you all know, this is a fairy tale and fairy tales sometimes have farfetched logic, i.e., a son today is a father tomorrow and a father today is going to be a grandfather tomorrow. This coined sentence will look a bit out of frame from the above ‘GFS‘ explanations. To explain this, let’s go over the details of the GFS methodology.
GFS is a backup methodology followed in storing your backup data by merging them on a ‘daily, weekly and monthly’ basis. This merge process is followed mainly to have multiple recovery points in times of restore and also to save space and reduce the incremental count in the chain of backed up data.
Let’s say you are configuring a disk image backup on an hourly basis, which means you will be having 24 backups per day, 168 backups per week and 5040 backups per month (in a 30 day month). This will not only look high numbered, but also have too many unnecessary restore points. Those unnecessary restore points also create storage issues. If a large file is being backed up in its 23rd incremental and modifications are saved at each incremental and the file is deleted while backing up the 455th incremental, its size will remain occupied until a merge is processed. The merge process will check for deleted files getting removed (if the settings are configured to remove deleted files) and will ensure duplicated blocks of data are removed, keeping one remaining. This will save a huge amount of space.
Also if a merge occurs on a periodic basis (daily, weekly or monthly), it will result in a reduced chain of incremental (about 7 files per week, 4 per month and finally with 12 per year). This in turn gives you a fixed number of necessary restore points. The merge process will also ensure a process of safeguarding your longtime saved data efficiently.
This clearly highlights the importance and benefits of the ‘GFS’ method implementation in a backup process.
We at Vembu, not only value our customers but also understand the importance that their backups have the ‘GFS’ method implemented in our disk image and VMware backup software. Currently we support merge processes to happen on an ‘hourly, daily and weekly’ basis in Vembu BDR v2.0, which ensures data restores in the easiest way possible. We also are working to have quarterly to half and year to year merges get added as a feature to ensure storage of data over longer periods for our customers. It would seem that ‘GFS’ backup worked out to have a ‘happily ever after’ after all.
