08-09-2012 05:14 AM
Hello everybody!!!
I have a Netbackup Catalog Police running in three diferents schedules - DIARY, HOURLY and WEEKLY.
My DIARY schedule starts at 09h AM and keep for more than 24 hours running and don´t finish. So, I have to kill it.
Does anybody see something similar in your NB Catalog ?
Thanks
Solved! Go to Solution.
08-10-2012 05:53 AM
Look at your schedules - seems Inline Tape Copy is configured. I would remove NFS destination as a start.
Please post your catalog policy config:
/usr/openv/netbackup/bin/admincmd/bppllist <policy-name> -U
bptm and bpbkar log files are correct ones, but you need to allow it to finish in order to find useful info.
The bpbkar test suggested by revaroo will start a local backup of the folder that you specify in the command and write the backup to /dev/null. This is how we test local read speed of the disk/filesystem where the folder is located.
The command will run in the background.
'ps -ef' or 'bpps' will show the bpbkar process.
Once finished, you can check result in bpbkar log.
See this TN for more info:
Measuring disk performance with bpbkar
http://www.symantec.com/docs/HOWTO56131
08-09-2012 05:52 AM
How big is the catalog ?
If it is large, this could be 'normal' .
If it is not large, then it could be simply poor backup performance.
To get the size, how big are the contents of these two directories (and everything under them)
/usr/openv/netbackup/bin
/usr/openv/db
Thanks,
Martin
08-09-2012 05:57 AM
mph, think you meant /usr/openv/netbackup/db not /usr/openv/netbackup/bin
08-09-2012 05:59 AM
Martin
These are my directories that you mentioned. Are you sure they contain my catalog?
-bash-3.00# cd /usr/openv/netbackup/bin
-bash-3.00# du -h .
1K ./bprd_child
1K ./bprd_parent
1.8M ./driver
2.5M ./goodies/support
948K ./goodies/man
96K ./goodies/misc/Linux/RedHat2.4
98K ./goodies/misc/Linux
518K ./goodies/misc/ALPHA/OSF1_V5
520K ./goodies/misc/ALPHA
644K ./goodies/misc/PC-IA64/WindowsXP
645K ./goodies/misc/PC-IA64
157K ./goodies/misc/RS6000/AIX5
157K ./goodies/misc/RS6000/AIX4.3.3
315K ./goodies/misc/RS6000
180K ./goodies/misc/SGI/IRIX65
182K ./goodies/misc/SGI
518K ./goodies/misc/PC/WindowsNT
520K ./goodies/misc/PC
104K ./goodies/misc/Solaris/Solaris7
106K ./goodies/misc/Solaris
98K ./goodies/misc/HP9000-800/HP-UX11.11
98K ./goodies/misc/HP9000-800/HP-UX11.00
198K ./goodies/misc/HP9000-800
2.5M ./goodies/misc
17K ./goodies/mediascripts
11M ./goodies
2K ./bprd.d
4K ./bpsched.d
12M ./support
188K ./private
101K ./cluster/suncluster
66K ./cluster/vcs
924K ./cluster/util
1.2M ./cluster
1K ./bpdbm_parent
116M ./admincmd
1K ./bpdbm_child
453M .
-bash-3.00# cd /usr/openv/db/
-bash-3.00# du -h .
1K ./staging
8K ./tix
132K ./scripts/wdata
4.1M ./scripts
1K ./shared
2.5M ./charsets/unicode
2.5M ./charsets
3.9M ./log
47K ./help/dbadm
49K ./help
37M ./bin
19M ./lib
192M ./data
1.0M ./res
260M .
08-09-2012 06:19 AM
Thanks for replying. Thats my /usr/openv/netbackup/db output.
How can I check if the Catalog backup is runnig ok?
And how can I see how much bytes has it backup up?
11K ./images/svapldesenv/1342000000/tmp
34M ./images/svapldesenv/1342000000/catstore
34M ./images/svapldesenv/1342000000
9K ./images/svapldesenv/1344000000/tmp/catstore
11K ./images/svapldesenv/1344000000/tmp
17M ./images/svapldesenv/1344000000/catstore
17M ./images/svapldesenv/1344000000
68M ./images/svapldesenv
97G ./images
08-09-2012 06:33 AM
Do you have bpbkar log (on the master) and bptm log (on the media server that is backing up the catalog) setup? If so, post the logs that match the time of the backup.
08-09-2012 06:46 AM
Here are the files....
08-09-2012 06:51 AM
bpbkar seems to have stopped here:
08:23:51.802 [27863] <4> bpbkar PrintFile: /opt/openv/netbackup/db/class/_svts2-files/schedule/
Run at the unix command prompt (logged in as root):
bpbkar -nocont /opt/openv/netbackup/db/class/_svts2-files/ > /dev/null 2>&1
It should return to the command prompt, if it doesn't check the bpbkar log, is it stuck on the file above?
08-09-2012 07:06 AM
That´s the output:
-bash-3.00# bpbkar -nocont /opt/openv/netbackup/db/class/_svts2-files/ > /dev/null 2>&1
-bash-3.00#
It returned to the prompt. What does this comand do ?
08-09-2012 07:49 AM
Opps, yes I did - thank you for correcting this
08-09-2012 08:33 AM
It destroys your catalog and policies...it'll run much faster now, just you see.
08-09-2012 08:37 AM
Only joking,,,it backs up that directory and below and throws the backup in the bin, ie it just read the data, a test to see if it can read the data.And it can.
08-09-2012 09:29 AM
Wonder if this is relevant - not quite the same, but closeish ...
http://www.symantec.com/docs/TECH71176
08-09-2012 01:55 PM
what speed does it run at? Running to tape or Disk ?
08-10-2012 05:07 AM
mph, thanks for the link.
Sajeev, my netbackup runs two copies. The first one in LTO2 tape and the second to a NFS mounted in the netbakup server.
Take a look in my rate transfer. Its ridiculous.
08-10-2012 05:21 AM
Have you tried to just run to one destination at a time?
with Inline tape copy backup speed will be limited to the speed of the slowest device.
It is probably the NFS destination that is so slow.
LTO2 destination (cannot believe these old devices are still in use) - locally attached to the master or to a media server?
08-10-2012 05:33 AM
Marianne
My master and media server are the same machine (Sun Fire V490).
How can I run the Catalog for just one destination?
Where are the Catalog logs that can show me if its running ok ?
Thanks a lot.
Andrich
08-10-2012 05:53 AM
Look at your schedules - seems Inline Tape Copy is configured. I would remove NFS destination as a start.
Please post your catalog policy config:
/usr/openv/netbackup/bin/admincmd/bppllist <policy-name> -U
bptm and bpbkar log files are correct ones, but you need to allow it to finish in order to find useful info.
The bpbkar test suggested by revaroo will start a local backup of the folder that you specify in the command and write the backup to /dev/null. This is how we test local read speed of the disk/filesystem where the folder is located.
The command will run in the background.
'ps -ef' or 'bpps' will show the bpbkar process.
Once finished, you can check result in bpbkar log.
See this TN for more info:
Measuring disk performance with bpbkar
http://www.symantec.com/docs/HOWTO56131
08-10-2012 06:16 AM
Thanks everybody
Reading all the posts, the correct decision is changing my LTO destination to a LTO4 and make the Catalog DISK copy to a local disk and not a NFS.
Andrich