@advnm,
Regarding option 1:
Forget about it. If it is not supported, don't even try it.
Regarding option 2:
The problem of accelerator is that it keep track of block changes in a localy (regular file) on the client. For each records, it will associate a checksum used for the next job to know what was changed since the last backup. If for some reason, the checksum does not match or the tracking log database is not accessible, it will scan the entire file system and run a full backup. this is what you are facing:
"'not using change journal data for <\\cifs\share\>:forcing rescan, each file will be read in order to validate the checksums.'"
As the checksum is not correct, it will rescan (force rescan) the entire volume and do a full backup, again and again untill the checksum is correct. the only solution for you is to do the following:
- Make sure no job are running for this client
- Stop netbackup client service
- delete All subfolders under c:\program files\veritas\netbackup\track
- restart netbackup client service
- run a new full backup for this client in order to rebuild the tracking log database
- run another full to see if it validate the checksum and send less Kb to the media server
About ACLs restore (see Netbackup 7.5 admin guide Vol I pg 119-121), they will be restored except under the following condition:
- cross-platform restore
- tar modification outside netbackup
To restore data without ACLs, the following condition must be met:
- MS-Windows Policy type
- Restore must be done from BAR GUI on netbackup server (windows or Unix)
- destination and source client must be running windows
Regarding option 3:
I did a lot of test with Accelerator, NT journaling and TIR (Tru Image for synthetic backup) and it depend on the type of data you need to backup. You can take in consideration the following:
- For fileserver with dynamics data for which there are a lot of changes (user home dir, shared document, packages ...) do a traditionnal backup without any of the above features (Accelerator, NT journaling or TIR) like this:
In the same Policy per Client:
=> 1 Weekly full scheduled
=> 1 daily differential/incremental scheduled
I do everyday backup of fileserver for ~30TB
- For Client with statics data (no regular changes like c:\, / ...) you can combine it with one of the above feature
But As your appliance is not supported by Symantec, you should export through NFS (or CIFS) your data and take regular backup from the target (DFS, NFS Client ....) but you will face a problem with ACLs because they are stored on the source which is not supported (example: on Solaris, NFS mounted resources with acls will not be restored. You need to take backup from the source to restore ACLs)
