Change Journal , track log , Accelerator and media side dedupe

[adamt078]

Hi guys ,

I am a newbie to netbackup and trying to understand a couple of things . I am in the process of backing up a server2012 client file server 3.5TB in size and have Media side dedupe congfigured  as client side not yet supported in Netbbackup 7.5.0.6

I however have track logs and change journal and accelerator enabled .

Now

1)  with Media side dedupe  enabled i guess that will send all the data (3.5TB) accross to the media server each time the backup runs and will then be deduped on the Media server ?? 

2) so what exactly will the  track logs and change journal and accelerator enabled do ?

3) what is the best configuration setting i should be using to get the data backed up correctly and in good time making use of all the ?

once again am a newbie:)

[Nicolai]

Client side dedupe and accelerator is two different thing.

1: You are right - but with accelerator enabled only entire changed files will be transferred. With client side dedupe only unique block within the file would be transferred.

2: To detect the changed files without reading the entire file system (potential million of files).

3:Using client side dedupe and accelerator, and specify ALL_LOCAL_DRIVES as file directive(VERY VERY high level).

[adamt078]

I cant however use client side dedupe in this case as not supported on server 2012

[adamt078]

so if I understand corectly with Media-side deduplication  and accelerator enabled ... From my understanding with accelerator and tracklogs enabled only the entire changed files will be transferred  only entire changed files will be transferred to the Media server where that will then be deuplicated .

But with

Client-side deduplication only only unique block within the file would be transferred . So really really with client -side deduplication the accerator plays no role ?

[Mark_Solutions]

You don't need client side de-dupe to use accelerator

The biggest difference will be your first "real" backup - after that there will not be a huge difference

The accelerator works at segment level so much deeper than being at file level - without client side de-dupe it will send all of the new segments to the media server but that will be way less than sending all changed files

Hope that makes sense

[adamt078]

Ok yes it does  Thank guys 

using change journal data for <K:\>: unable to validate change journal usage <reason=previous backup wasn't a successful backup>

what happens if I get  the message above as my backup seems to be doing the entire 3.5TB ? Although running faster than the first "real backup"

[Mark_Solutions · Accepted solution]

If it cannot use the change journal it will still use the track log if it can

To corrected it and get even better rates again you would need to run a forced re-scan schedule again  - that being a "real" full backup which will obviously take much longer

If it wasn't using accelerator it would tell you so in the details of the job so it looks like it is still using it but wont be as fast as when CJ is also available

[adamt078]

Thanks for your help Mark  I am also getting this below 

NOT using change journal data for <K\>: hard link or reparse point change detected

After reading this article 

https://www-secure.symantec.com/connect/forums/accelerator-cant-use-journal-hard-link-or-reparse-point-change-detected

It shows all my accelerator backups are just the tracklogs ONLY .... How can i get the change journal working  as this will be the faster way i think 

[Mark_Solutions]

I dont think it can be done - it is a MS issue as covered in the post linked by yourself