Change journal enabled, but "hard disk or reparse point change detected"

[Georges_N]

Hello

We've just recently updated our Netbackup from 7.5.0.6 to .7 and also recently started using the Journal feature for the first time (with our new NBU appliance) - I haven't had issues with it until this particular file server (different policies for each large disk):

2/5/2014 9:02:35 PM - Info bpbkar(pid=1528) change journal enabled for <D:\>      

2/5/2014 9:03:26 PM - Info bpbkar(pid=1528) NOT using change journal data for <D:\>: hard link or reparse point change detected

2/5/2014 9:02:16 PM - Info bpbkar(pid=4960) change journal enabled for <J:\>        

2/5/2014 9:02:49 PM - Info bpbkar(pid=4960) NOT using change journal data for <J:\>: hard link or reparse point change detected

2/5/2014 11:01:22 PM - Info bpbkar(pid=6068) change journal enabled for <G:\>        

2/5/2014 11:01:42 PM - Info bpbkar(pid=6068) NOT using change journal data for <G:\>: hard link or reparse point change detected

This makes the daily Diff-Inc backups extremely long (up to 15 hours). I've looked at other posts with the same issue but they seem to have this error after a certain time, due to timeout. This error, for us, occurs at the beginning of the backup. We get this message and then it proceeds to basically rescan the whole disk (the 15-hour job was on an 8tb disk)

Any ideas? Thank you.

Here's some info:

Symantec Netbackup 7.5.0.7 running on Windows Server 2008 R2 x64 master server

Symantec Netbackup Appliance 5230 running version 2.5.4

The File Server is running Windows 2008 R2 x64 as well, with 7 RDM disks ranging from 8TB to 0.5TB

This error hasn't happened anywhere else, as far as I can tell (out of like 150 clients or more)

[Georges_N]

Bonus information:

We also use Sympana Commvault for Archiving (only) - it uses a Stub system which may interfere... if this is the problem, any way for both to coexist?

[Mark_Solutions · Accepted solution]

No way around it unfortunately for that client - at least not as far as the journal goes

As you have an appliance you are running de-dupe so just use accelerator to do the backups (with client side if that client can handle the processing - but not essential)

It wil help a little with the dailt incrementals due to the track log - not hugely but a bit - but you will need to change that client to useing archive bit or time stamp for incrementals.

Probably a better way would be to split that server into more streams that run at the same time to speed things up - or use flashbackup which would be the fastest way when it comes to incrementals

In any case the change journal cannot track file changes due to the stub files so when NetBackup queries it the change journal just tells netbackup it does not know what has changed!

Do also make sure that bpfis and bpbkar32 are excluded from recalling files on the server - otherwise it may be slow due to recalling everything during the backup

Hope this helps

A few bits of info here : http://www.symantec.com/business/support/index?page=content&pmv=print&viewlocale=&id=HOWTO68492

[Georges_N]

Thanks Mark, not the answer I wanted to read, but an answer nonetheless.

We used to run FlashBackup-Windows but accelerator seems to do the trick via MS-Windows. I'll try another incremental right now without the Journal.