gibbsie.org Knowledge Base

Veritas NetBackup

Aug 3rd 2008
No Comments
respond
trackback

Calculating Total Data and the Number of Tapes

Total Data = Size of full backups + Size of incremental backups

Size of full backups = (Amount of data to back up) * (Frequency of backups) * (Retention period)

Size of incremental backups = (Churn Rate Percentage * Amount of data to back up) * (Frequency of backups) * (Retention period)

Number of tapes required = (Total Data) / (Tape capacity)

Example: Calculating backup sizes

Preliminary calculations:

Size of full backups = 500 GB * 4 (per month) * 6 months = 12 TB

Size of incremental backups = (20% * 500 GB) * 30 * 1 month = 3 TB

Total data = 12 TB + 3 TB = 15 TB

Veritas NetBackup Version 6

Controlling NetBackup Processes

Starting and stopping NetBackup:

# /etc/init.d/netbackup stop

# /etc/init.d/netbackup start

Check to see if any processes remain:
# /usr/openv/netbackup/bin/bpps -a

If any remain, ensure all processes are killed:
# /usr/openv/netbackup/bin/goodies/bp.killall

To restart just the EMM database, which is required if all disk space was consumed and has since had space created, check that the local host is the EMM server:

# /usr/openv/netbackup/bin/bpclntcmd -is_local_host servername

Then restart if the localhost is the EMM server:

# /usr/openv/netbackup/bin/nbdbms_start_stop stop

# /usr/openv/netbackup/bin/nbdbms_start_stop start

Unified Logging

Enable logging:

The following enables the maximum level of logging for the ndmpagent and ndmp processes:

# /usr/openv/netbackup/bin/vxlogcfg -a -p 51216 -o 134 -s DebugLevel=6

# /usr/openv/netbackup/bin/vxlogcfg -a -p 51216 -o 151 -s DebugLevel=6

Retrieve logs:

The following extracts all captured logs to a file. Note: this can be time consuming!
# /usr/openv/netbackup/bin/vxlogview -i ndmpagent -d T,s,x,p > /tmp/ndmpagent.out &

# /usr/openv/netbackup/bin/vxlogview -i ndmp -d T,s,x,p > /tmp/ndmp.out &

NDMP

Verify NDMP communications:

/usr/openv/volmgr/bin/set_ndmp_attr -verify hostname.domainname

NDMP Progress Timeout: (should be 1440, which is 24 hours)

cat /usr/openv/netbackup/db/config/NDMP_PROGRESS_TIMEOUT

If multiple interfaces exist on the media server, specify the correct hostname to announce via NDMP:

echo "NDMP_DATA_CONNECTION_HOST_NAME=hostname" >> /usr/openv/netbackup/db/config/ndmp.cfg

Where hostname is either IP or the hostname as a string.

NDMP Diagnosis:
These examples assume the NDMP client is a NetApp filer. First, check that “options ndmpd.access” has the IP address of the NetBackup media server that will service this filer:
options ndmpd.access host=192.168.123.234

If multiple interfaces exist on a NetApp, ensure a prefered interface is set:
options ndmp.preferred_interface interface_name
(where the interface name might be e0, etc)

Force the ndmpd process to restart on a NetApp:
ndmpd off
ndmpd on

Set ndmpd debugging to maximum:
ndmpd debug 50
(Logs to /etc/log/ndmplogs* on a NetApp filer as well as outputted to the console)

Turn off ndmpd debugging using:
ndmpd debug 0

Controlling Backup Selection using Exclude and Include Lists

Distributing an exclude_list to a list of NetBackup clients
# for THOST in `cat filelist`; do
> /usr/openv/netbackup/bin/bpgp to $THOST /tmp/exclude_list /usr/openv/netbackup/exclude_list
> done

The general syntax for using the undocumented bpgp command is:

bpgp from client filename_on_client filename_on_server
bpgp to client filename_on_server filename_on_client

Veritas NetBackup Version 5

Controlling NetBackup Processes

Starting and stopping NetBackup:

# /etc/init.d/netbackup stop

# /etc/init.d/netbackup start

Check to see if any processes remain:
# /usr/openv/netbackup/bin/bpps -a

If any remain, ensure all processes are killed:
# /usr/openv/netbackup/bin/goodies/bp.killall


This post is tagged , ,

No Comments

Leave a Reply