This is a result of a customer ticket correspondence, where jscott explains how does disksync work, what are rolling in and out of the data, what does retention mean and how it can and should be set, what are safesets, what are sync jobs and so forth. i think it might help people using disksync :

----------------------------------------------------------------------------------------
(jscott-03/16/2005 17:16:25):

Greetings,

One safeset is created each time you run your task. A task is typically scheduled to run once per day. A safeset and a "copy" are the same thing.

Seven days of backups, running your task once per day, equals seven copies. A copy equals a safeset, and a safeset is everything that has changed since the last safeset was taken.

If you have a clean disksync account, and 13gb of data that changes by 1gb daily, you will have one safeset that is 13gb, and 6 safesets that are 1gb. On a typical one-week (7 days AND 7 copies) retention, on the 8th day, when the 8th copy is taken, the oldest data is "rolled up" into the seed, the oldest copy is destroyed, and the process repeats.

A Synchronize Task is NOT a Backup Task. Only a Backup Task creates a backup. A Synchronize task merely synchronizes the settings on your Agent with the settings that are stored on the Vault, and is only necessary if you re-install your DiskSync Agent and Re-register as an already registered computer.

When you say "disksync does a seed backup of the whole server", this is only true if your Task Source includes the whole server in the Task. Each subsequent Backup Task operation takes the deltized data into a safeset, and at the end of your retention period, if the retention criteria are met, the oldest safeset is rolled up or rolled out of the seed, depending on several factors.

In the case of my example of a 13gb Task Source, your initial seed will consume 13gb of space (not considering compression) on the vault, and if 1gb of that Source data changes daily, each subsequent Backup Task operation will store an additional 1gb to the vault. This will consume 19gb on the Vault after one week. If the retention period is 7 days and 7 copies, the 8th day's copy will either roll up or roll out 1gb from the seed, increasing or decreasing the overall storage utilization by 1gb.

Definitions:

"roll up" - When new data is added to your server, and the retention criteria are met, if this data still exists on your server, it will be required if you need to restore the data. This data is "rolled up", or added to the seed data.

"roll out" - When new data is collected by DiskSync, but later is deleted from your server, this data will "roll out", or be removed from, the seed once the retention criteria have been met for the date the data was deleted.

For example, with a 7/7 retention:

(Compression not considered, because it's easier to see how this works without considering the compression)
Day 1: Safeset 1: 13gb collected
Vault Pool = ~13gb
Day 2: Safeset 2: 1gb new data collected
Vault Pool = ~14gb
Day 3: Safeset 3: 1gb new data collected
Vault Pool = ~15gb
Day 4: Safeset 4: 1gb new data collected, Day 2's data deleted from your server
Vault Pool = ~16gb
Day 5: Safeset 5: 1gb new data collected, Day 3's data deleted from your server.
Vault Pool = ~17gb
Day 6: Safeset 6: 1gb new data collected, Day 4's data deleted from your server.
Vault Pool = ~18gb
Day 7: Safeset 7: 1gb new data collected, Day 5's data deleted from your server.
Vault Pool = ~19gb
Day 8: Safeset 8: 1gb new data collected, Day 6's data deleted from your server, Day 1's data rolls up into the seed.
Vault Pool = ~20gb
Day 9: Safeset 9: 1gb new data collected, Day 7's data deleted from your server, Day 2's data rolls up into the seed.
Vault Pool = ~21gb
Day 10: Safeset 10: 1gb new data collected, Day 8's data deleted from your server, Day 3's data rolls up into the seed.
Vault Pool = ~22gb
Day 11: Safeset 11: 1gb new data collected, Day 9's data deleted from your server, Day 4's data rolls up into the seed, Day 2's data rolls out of the seed.
Vault Pool = ~22gb
Day 12: Safeset 12: 1gb new data collected, Day 10's data deleted from your server, Day 5's data rolls up into the seed, Day 3's data rolls out of the seed.
Vault Pool = ~22gb
Day 14: Safeset 14: 1gb new data collected, Day 11's data deleted from your server, Day 6's data rolls up into the seed, Day 4's data rolls out of the seed.
Vault Pool = ~22gb

As you can see in this example, since the data from Day 2 was deleted from your server on Day 4, after 7 days have passed, the data is removed from the backup because it has expired. The backup will continue to grow so long as the data set in the Task Source grows. If the data in the Task Source is deleted from your server, at the end of the retention policy, the data will be rolled out of the system.

This gives you 7 "points in time" which you can use to restore the deleted data.

As you can also see from this example, on Day 10, the Vault Pool aproximately totals 22gb, and on Day 11 it remains at 22gb because 1gb of new data is added daily, and 1gb of data is removed from your server daily.

This is how a retention policy works on DiskSync.

If you make one backup a week, and you are comfortable with eliminating six points-in-time throughout the week from your backup strategy, and you only wish to keep one week's backup on hand at any given time, you can change your retention policy to include 1 day and 1 copy. The next time a backup is taken, even if it is a manual copy, that backup becomes the only backup available for the system.

Sincerely,
Justin Scott
Managed Services Engineer
The Planet
----------------------------------------------------------------------------------------