Created Date: 3/14/2018
Last Updated: 3/14/2018
Description: See README.md for complete details. This contains the files corresponding to KB #2357 which are used as an example of using the synbackup utility.
Platforms: Unix
Products: Synergy DBL; Synergy DBMS
Minimum Version: 10.1.1
Author: Galen Carpenter
Additional Information: Using synbackup on Unix or Linux
When backing up Synergy databases, problems can arise if a file is in the process of being updated when a backup occurs, which may result in a corrupted file. To help solve this problem, you can use the Synergy/DE synbackup utility to freeze the updating of the files, allowing them to be backed up without corruption. After the backup is completed, the synbackup utility can thaw the I/O and allow updates to the files again. This freeze/thaw process is called quiesce.
This CodeExchange entry contains several script files to help you automate the use of synbackup on UNIX. These files should work with a variety of backup applications.
stop_APP.sh and start_APP.sh. The *_APP scripts do the actual freezing
and thawing of the Synergy I/O by calling the synbackup utility.
The *_APP.sh scripts need to be modified to specify the location of
the setsde script in your Synergy distribution directory. They
generate log files (stop_APP.log and start_APP.log) that show the time
they were run and the output from the synbackup command.
synbackup (or K30synbackup and S30synbackup for AIX). These scripts
execute the synbackup -c command to create the shared memory that
informs all Synergy programs that I/O is enabled or disabled. Look at
the script(s) to verify that DBLDIR is set correctly and the path for
synbackup is correct. If you have multiple installations of Synergy on
the machine, see "If you have multiple installations of Synergy" below.
The synbackup (or S30synbackup) script should be run automatically
every time the system starts up. (K30synbackup is run on shutdown to
do synbackup -d.) See "Running the synbackup script at startup" below
for instructions on the various platforms.
The other script file you need to be concerned with is setsde, which is included in your Synergy distribution. To initialize the use of the synbackup utility, uncomment the setting of the SYNBACKUP environment variable in setsde. (SYNBACKUP must be set before the synbackup -c command is run.)
How it works:
When the system is booted, the synbackup (or S30synbackup on AIX) script is run to set the SYNBACKUP environment variable "on" and run the synbackup -c command, which creates the shared memory.
When a backup is requested, the backup software calls stop_APP.sh, which runs synbackup -b and then sleeps 10 seconds (to allow the program time to finish up what it is doing). Then stop_APP calls synbackup -s to freeze the I/O.
When the backup finished, the backup software calls start_APP.sh to run synbackup -x to unfreeze the I/O.
If you have multiple installations of Synergy:
When the SYNBACKUP environment variable is set, the runtime looks for the file DBLDIR:synbackup.cfg, which is created by the synbackup -c command. If there are multiple installations of Synergy on the system, each will have DBLDIR set to its own installation area. To have all of the installations use the same shared memory, you need to create a symbolic link to synbackup.cfg in the DBLDIR directory of each additional installation location. For example: "ln -s /usr2/test_1033_32/synergyde/dbl/synbackup.cfg /usr2/test_1033d_32/synergyde/dbl/synbackup.cfg". Be sure to enable the SYNBACKUP environment variable in the setsde script for each synergy installation area.
Running the synbackup script at startup:
Ubuntu: Copy the synbackup script file to /etc/initd and make sure it
is owned by root with permissions of 777. Then do
"/usr/lib/insserv/insserv synbackup" to have the system create the
appropriate links to the /etc/rc.d/rcN.d files.
Redhat: Copy the synbackup script file to /etc/initd and make sure it
is owned by root with permissions of 777. Then do
"chkconfig --add synbackup" (that is, dash dash add) and
"chkconfig --level 2345 synbackup on" (that is, dash dash level).
The chkconfig command may be located in /sbin.
AIX: Copy the K30synbackup and S30synbackup scripts to the
/etc/rc.d/rc2.d directory and modify them for the location of your
Synergy installation. The S30synbackup script is for startup and the
K30synbackup script is for shutdown. Both have the same contents. Be
sure they are owned by root/system with file permissions of 640.
Backup software specific instructions:
These three commonly used backup applications can be modified to work with the stop_APP.sh and start_APP.sh scripts.
Semantec NetBackup:
NetBackup looks for two files, bpstart_notify and bpend_notify, in the
/usr/openv/netbackup/bin directory. If it finds them, it executes
bpstart_notify before the backup starts and bpend_notify after the
backup has completed. The CodeExchange zip file includes versions of
these scripts that we downloaded from
https://vox.veritas.com/t5/NetBackup/Pre-Post-script-commands-for-backup/td-p/762844.
The bpstart_notify script calls stop_APP.sh and the bpend_notify scipt
calls start_APP.sh. Set APP_SCRIPT_LOCATION in the bp*_notify scripts
to point to the location of start_APP.sh and stop_APP.sh. (Because
synbackup must be run as root, the *_APP scripts are started directly
by the bash command rather than going through another account.)
DOLOGGING is set to 1 in the scripts and should be turned off once you
are done testing and everything is running successfully. We recommend
you read through these scripts to see what else they do.
Veeam Backup & Replication:
In the "Processing Settings", on the "Scripts" tab for
"Guest Processing", the location and name of the Pre-freeze and
Post-thaw scripts need to be filled in with the location of the
stop_APP.sh and start_APP.sh scripts, respectively.
Microlite's BackupEdge:
In the "Backup Domain" under "Advanced Properties" there is a place to
specify a "Start/Stop Script". The default is
/usr/lib/edge/bin/edge.bscript, which is run before and after the
backup is performed as well as before and after verification of the
backup. Microlite does not recommend modifying this script. At the
start of a backup, edge.bscript runs the /etc/edge.start script and at
the conclusion of a successful backup and verify, it runs the
/etc/edge.passed script. If the backup or verify fails, the
/etc/edge.failed script is run. These three scripts are user
modifiable. You can modify /etc/edge.start to run the stop_APP.sh
script, and then modify /etc/edge.passed and /etc/edge.failed to run
the start_APP.sh script. (New releases of BackupEdge replace these
three scripts and rename the existing files by appending 00 to the
name, so you won't lose your changes.)
If you are using other backup software you may still be able to use the *_APP scripts. Most backup software has a way to specify freezing and thawing of I/O. Search your backup software's documentation for "quiesce" or "freeze" or "thaw" to find out how it is specified. Then configure it to call the stop_APP.sh script to freeze the file writes and the start_APP.sh script to thaw the file writes.
(This information is included in KB article 2357.)