Mystery solved - or so it seems.
Last week we had a helpdesk ticket where the user wanted some chmod/chgrp
modifications of all files/directories in this area.
My co-worker did that on Thursday/Friday of last week. Seems this changes
the ChangeTime (ctime), which NBU is using for determining if something has
"Changed". I then did a Differential, and it was looking at the ctime.
For one of the files:
stat error.txt
File: `error.txt'
Size: 338713 Blocks: 962 IO Block: 4096 regular file
Device: 1dh/29d Inode: 4365470520 Links: 1
Access: (0664/-rw-rw-r--) Uid: ( 2370/ cgoina) Gid: ( 9003/jtc-solexa)
Access: 2008-10-15 09:44:41.235029187 -0400
Modify: 2008-10-15 09:44:41.235029187 -0400
Change: 2008-12-09 19:00:21.895757388 -0500
Turns out, the chmod/chgrp needed to be rerun on Monday/Tuesday this week.
Is there a way for NBU to look at mtime instead of ctime?
Thanks,
Randy
----- Original Message -----
From: "A Darren Dunham" <ddunham AT taos DOT com>
To: <Veritas-bu AT mailman.eng.auburn DOT edu>
Sent: Wednesday, December 10, 2008 11:23 AM
Subject: Re: [Veritas-bu] Full backup followed by a Differential = Another
Full
> On Wed, Dec 10, 2008 at 07:42:41AM -0800, Randy Doering wrote:
>> Backup Client = Isilon NAS Unit - NDMP. NBU version is 6.5.3.
>
>> With this?one (10+TBs, 1,643,099 files) during the backup NBU said it
>> was doing a Differential, but when it was all done, it had actually
>> backedup everything again.
>
>> If I kick off another Differential this coming weekend, I certainly
>> don't want to have yet another backup that gets all of the data once
>> again.
>>
>> Any thoughts?
>
> Since this is NDMP, Netbackup isn't making any of the file level
> decisions. It's just passing a set of parameters to the host (dump
> level, date of backup, file path, tape drive, etc...) and letting the
> make the selections.
>
> I don't have it in my notes, but you can up the log level of ndmp on the
> isilon and view exactly what NBU is sending over. It'll show up in the
> ndmp logs. Ah. Found my notes...
>
> On any node of the cluster, edit the file /etc/mcp/templates/syslog.conf
> and change the line:
> *.>=info /var/log/isi_ndmp_d
> to:
> *.* /var/log/isi_ndmp_d
> This should increase the logging levels on the cluster for ndmp.
>
> You may want to check with your support contacts before doing this, but
> it's just a log level change.
>
> With that in place, you can start a manual differential and see what it
> looks to be doing. You'll be able to kill it so it won't consume tapes
> for you. If nothing looks obviously wrong, you may have to contact
> support.
>
> Good luck!
>
> --
> Darren
> _______________________________________________
> Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
|