• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 474
  • Last Modified:

How exactly do backup snapshots work?????

Our company is getting ready to implement a snap server 4500 and Symantec's LiveState Recovery backup software.  I understand what a snapshot does, but how exactly does it work.  If I am going to take a snapshot of a system volume that I will send over the network to the snap server for backup how does the snapshot exactly work.  If I am correct a snapshot only takes a few seconds to create.  If we bring down a database for a minute to have the data snapped then the database returns to it's live state what stops a newer version of a file from being backuped and not the snapped verison.  Also where is the work being done to create the actual full snapshot file (IE. the server or backup device).  I hope that made some sense.  Thanks for your help.  
0
icarus2256
Asked:
icarus2256
  • 2
1 Solution
 
CaseybeaCommented:
I can't speak for SYmantec's software-- but in general, here is how snapshots work- I hope this helps:

When a snapshot is taken, basically that is the moment the "clock" starts ticking.   From that moment forward, the system starts "logging" the BEFORE-STATE changes of data.

I'll use a simple example of a file.   The file is "fred.txt" and contains three lines:

moo
cow
bird

Now, we take a SNAPSHOT.

Then, some time after the snapshot--  we EDIT that file, and change it.   (it now says "moo, cow, DOG").    BUT---  the snapshot software, that's been running since we started it a while back, REMEMBERS that the last line of that file was "bird".

From a storage perspective-- the only "space" utilized by a snapshot is the space required to log (what was there before) changes were made.  

Physically, something like 95% of the snapshot data is actually copied from the LIVE RUNNING VOLUME.

It's pretty cool technology.

It's also much different and more complex than that; the above is a simplified example.   Most snapshot technology actually works from block-level of storage volumes, and not "parts of files".   It actually monitors "parts of disks", and when something changes on the live running volume-- the snapshot keeps a copy of what was there BEFORE.....

If you think about this a bit more- you also have to know that the "longer" you leave a snapshot in place--  if the volume you've "copied" changes a lot, the space required for the snapshot information will grow and grow and grow,....    ultimately, you NEED TO kill (remove) the snapshot data.

I hope this exlanation helps.    Remember- it's more complicated than this; I am simply trying to put it in terms that makes sense for most folks.

-kcb
0
 
icarus2256Author Commented:
So let make sure I have this right.  If I am backing up a file and the contents of the file include three lines showing:

moo
cow
bird

Then at that moment I take the snapshot.

Then, while the data is being backuped someone changes the file and adds the line:

dog

What if the backup job was complete before the file was changed.  The file I backuped would say (moo, cow, bird) correct??  So basically a snapshot is like a base full backup with incrementals after that point??  Please clarify.  Thanks
0
 
icarus2256Author Commented:
My IT manager just asked me a great question.  Let's say I backup an entire database and then continue forward with incremental snapshots.  Let's also say a table inside of that database contains 2,000,000 records and one day just a single record is changed.  Now is my snapshot going to copy across just that single record change or the entire database with 2,000,000 records??  What kind of file extension would that single record change be in??  I basically want to know how fast disk space could be used up with the snapshot.  Thanks
0
 
WeHeCommented:
It will only copy the blocks of information on your disks which have changed.
But keep in mind, the worst case for a snapshot is a complete data change, what means you need space for the snapshot in the same size as the active volume.
In your example it will be only the harddisk space, the single record needs.
btw, a incremental snapshot will need more space over time, because every snapshot will grow with every datachange.
> What kind of file extension would that single record change be in??
You can use the snapshot like the original filesystem. You will see the databasefile as normal, just the record will be the old.
>  I basically want to know how fast disk space could be used up with the snapshot
It depends on your data changing rate.
If you change about 3GB Data per day, every snapshot will grow about 3GB per day.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

Join & Write a Comment

Featured Post

The 14th Annual Expert Award Winners

The results are in! Meet the top members of our 2017 Expert Awards. Congratulations to all who qualified!

  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now