Does this help then ? =) http://seer.support.veritas.com/docs/236274.htm
Talks about the CLEAN_IN_BACKGROUND file and bpimage -cleanup. Now you have
something you can point at.
Nice write-up btw Dan.
John Nardello
T-Mobile: Enterprise Systems Backup Group Technical Lead
"Your backups are only as good as your restores."
-----Original Message-----
From: veritas-bu-admin AT mailman.eng.auburn DOT edu [mailto:veritas-bu-admin
AT mailman.eng.auburn DOT edu]
Sent: Monday, January 31, 2005 7:17 AM
To: 'veritas-bu AT mailman.eng.auburn DOT edu'
Subject: RE: [Veritas-bu] Non-destructive ways to reduce the size of the
Netbackup database ?
We already have the CLEAN_IN_BACKGROUND link. We already compress our images.
We have few, if any, "infinity" retentions.
It appears we're doing everything "right." Nearly all of our retentions are two
weeks or two months, but there are thousands and not enough time in the day to
inventory them all.
I am unable to find documentation on the "bpimage" command, particularly with
respect to this "-cleanup" flag. Without official documentation as to what the
below command does, from Veritas, I cannot run it to see if it will help.
-----Original Message-----
From: Dan Otto [mailto:Dan.Otto AT veritas DOT com]
Sent: Tuesday, January 25, 2005 4:51 PM
To: Kirsch, Mathew (Matt)
Subject: RE: [Veritas-bu] Non-destructive ways to reduce the size of the
Netbackup database ?
Before you start adding disk space, make sure the clean up is running thus
removing expired images-
Add this touch file-
# touch /usr/openv/netbackup/bin/CLEAN_IN_BACKGROUND
Run this in cron or manually and it should free any disk space taken up by
expired images.
# /usr/openv/netbackup/bin/admincmd/bpimage -cleanup -allclients
If you use retention periods of infinity, then look into catalog archiving
available in 4.5 FP3.
How NetBackup Database Cleanup happens.
The database clean up process is kicked off after every "scheduled run of
backups".
Definition of "scheduled run of backups" is the following:
1. The bpsched process is the process that "kicks off" backups.
2. Every 10 minutes (default value) bpsched wakes up and says "Is there
anything for me to do?"
3. If there is a backup scheduled, then bpsched creates a "work list". This
bpsched process is considered the "main bpsched"
4. When a work list is created, backups start happening. Main bpsched spawns
new bpsched processes to handle each job
in the work list
5. After 10 minutes, another bpsched process wakes up and says, are there more
backups for me to do?
6. If yes, that new bpsched process appends the found backup jobs to the main
bpsched process and then "goes away"
7. After another 10 minutes, another bpsched process wakes up and says "are
there more backups to do?"
8. If yes, # 6 happens again.
9. If no, the new bpsched goes away. When the main bpsched process finishes, it
then says "since I have no more backups
to run, it is time to clean the NetBackup database and it does so. The status
of this shows up in the All Log Entries
report.
Basically, by default, the cleanup process won't run if you have backups
running. If you extrapolate the above
and see that bpsched is constantly busy doing backups, then you see that it has
no free time to do the database
clean up. Thus, the clean up doesn't happen.
If there is nothing you can do about how busy you keep the bpsched process (you
must to your backups as scheduled), then
you can get the database clean up to run during your live backups.
# touch /usr/openv/netbackup/bin/CLEAN_IN_BACKGROUND
# /usr/openv/netbackup/bin/admincmd/bpimage -cleanup -allclients
That should tell the bpimage process to go ahead and clean up the database even
though NetBackup backups are still running
-----Original Message-----
From: veritas-bu-admin AT mailman.eng.auburn DOT edu [mailto:veritas-bu-admin
AT mailman.eng.auburn DOT edu] On Behalf Of Kirsch, Mathew (Matt)
Sent: Tuesday, January 25, 2005 11:41 AM
To: 'veritas-bu AT mailman.eng.auburn DOT edu'
Subject: [Veritas-bu] Non-destructive ways to reduce the size of the Netbackup
database ?
The volume on which we store our Netbackup database is edging closer and closer
to being completely full each week. Regular Netbackup database backups helps
some, but it's always a little bigger every day.
At the rate it's going, we're going to run critically short on space in the
next few weeks. We can expand the volume, but would rather not, because then
it's going to exceed the capacity of 1 tape.
Potentially what kinds of stuff can I clean up? Are there any settings I can
change, like on the retention of indexes or something?
|