Here's my full bash script:
#!/bin/bash
logs="$HOME/sitedb_backups/log"
mysql_user="user"
mysql_password="pass"
mysql=/usr/bin/mysql
mysqldump=/usr/bin/mysqldump
tbackups="$HOME/sitedb_backups/today"
ybackups="$HOME/sitedb_backups/yesterday"
echo "`date`" > $logs/backups.log
rm $ybackups/* >> $logs/backups.log
mv $tbackups/* $ybackups/ >> $logs/backups.log
databases=`$mysql --user=$mysql_user -p$mysql_password -e "SHOW DATABASES;" | grep -Ev "(Database|information_schema)"`
for db in $databases ; do
$mysqldump --force --opt --user=$mysql_user -p$mysql_password --databases $db | gzip > "$tbackups/$db.gz"
echo -e "\r\nBackup of $db successfull" >> $logs/backups.log
done
mail -s "Your DB backups is ready!" yourmail#gmail.com <<< "Today: "`date`"
DB backups of every site is ready."
exit 0
Problem is when i try to import it with mysql i am gettint error 1044 error connecting to oldname_db. When i opened sql file i have noticed on the first line CREATE command so it tries to create that database with the old name. How can i solve that problem?
SOLVED.
Using --databases parameter in my case is not necessary and because of --databases it was generating CREATE and USE action in the beginning of the sql file, hope it helps somebody else.
Use the --no-create-db option of mysqldump.
From man mysqldump:
--no-create-db, -n
This option suppresses the CREATE DATABASE statements that are
otherwise included in the output if the --databases or --all-databases
option is given.
Related
I have two tasks in task group
1) a db task to bring up a db and
2) the app that needs the db to be up.
Both start in parallel and the db tasks takes a lil bit time but by then the app recognizes that db is not up and kills the db task. Any solutions? Please advise.
It's somewhat common to have an entrypoint script that checks if the db is healthy. Here's a script i've used before:
#!/bin/sh
set -e
cmd="$*"
postgres_ready() {
if test -z "${NO_DB}"
then
PGPASSWORD="${RDS_PASSWORD}" psql -h "${RDS_HOSTNAME}" -U "${RDS_USERNAME}" -d "${RDS_DB_NAME}" -c '\l'
return $?
else
echo "NO_DB Postgres will pretend to be up"
return 0
fi
}
until postgres_ready
do
>&2 echo "Postgres is unavailable - sleeping"
sleep 1
done
>&2 echo "Postgres is up - continuing..."
exec "${cmd}"
You could save it as entrypoint.sh and run it with your application start script as the argument. eg: entrypoint.sh python main.py
I have set up a cron in my godaddy server for taking DB backup. For Testing purpose, I run the cron in every minute. The command is :
mysqldump tuniv_results > /home/username/public_html/DB-VVS/tuniv_results.sql
In my DB-VVS folder one file, tuniv_results.sql, is creating but it is of zero byte. Could you please let me know the issue, why it is not creating properly?
Thanks in advance.
------------UPDATE-------------------
$user="****";
$password="****";
$database="*****";
$dumpCommand='/usr/bin/mysqldump';
$dumpCommand.=" -e -f -h <ipaddress> -u$user -p$password";
$dumpCommand.=" $database";
$dumpCommand.=" > bekap.sql";
$results=$dumpCommand;
exec($dumpCommand);
echo "result: ".$results;
I create a file in the root folder and put the absolute path of that file in the Command text-field as /home/username/cronfile.php. But in the root file there is no file like bekap.sql. Please let me know what might be the issue.
Try this one :
$user="*********";
$password="*****";
$database="*********";
$dumpCommand='/usr/bin/mysqldump';
$dumpCommand.=" -e -f -h host.name.com -u$user -p$password";
$dumpCommand.=" $database";
$dumpCommand.=" > bekap.sql";
$results=$dumpCommand;
exec($dumpCommand);
echo "result: ".$results;
Another solution :
I think this will help to you.
Open terminal and type:
sudo tcsh
pico /etc/crontab
or
nano /etc/crontab
And add one of the following lines depending on your situation. This schedule the backup on 1am every day.
Remote Host Backup with linked PATH to mysqldump:
0 1 * * * mysqldump -h mysql.host.com -uusername -ppassword --opt database > /path/to/directory/filename.sql
Remote Host Backup:
0 1 * * * /usr/local/mysql/bin/mysqldump -h mysql.host.com -uusername -ppassword --opt database > /path/to/directory/filename.sql
Local Host mysql Backup:
0 1 * * * /usr/local/mysql/bin/mysqldump -uroot -ppassword --opt database > /path/to/directory/filename.sql
(There is no space between the -p and password or -u and username - replace root with a correct database username.)
Every week, I have to run a script that truncates a bunch of tables. Then I use the export data task to move the data to another server (same database name).
The servers aren't linked, I can't save the export job, and my permissions/settings are limited by the DBA (I am an admin on the databases). I have windows authentication on both servers only. The servers are different versions (2005/2008).
My question is is there a way to automate this with my limited ability to modify the servers? Perhaps using Powershell?
Selecting all these tables and stuff in the export wizard week after week is a pain.
If you have access to the SQL Server Management console apps, try something this from a different system.
C:\> bcp ExportImportFile.inp out prod.dbo.[Table] -b 10000 -S %SQLSERVER% -U %USERNAME% -P %PASSWORD% -T -c > C:\Temp\ExportImport.log
C:\> sqlcmd -S %SQLSERVER% -U %USERNAEM% -P %PASSWORD% -Q "Use Prod;TRUNCATE TABLE [Table];" >> C:\Temp\ExportImport.log
C:\> bcp prod.dbo.[Table] in ExportImportFile.inp -b 10000 -S %SQLSERVER% -U %USERNAME% -P %PASSWORD% -T -c >> C:\Temp\ExportImport.log
You can use DBATOOLS:
$splat = #{
SqlInstance = '{source instance}'
Database = 'tempdb'
Destination = '{dest instance}'
DestinationDatabase = 'tempdb'
Table = 'table1' # you can provide a list of tables
AutoCreateTable = $true
Truncate = $true
}
Copy-DbaDbTableData #splat
If you don't have dbatools: https://dbatools.io/getting-started/
I'm running the Microsoft SQLCMD tool for Linux (CTP 11.0.1720.0) on a Linux box (Red Hat Enterprise Server 5.3 tikanga) with Korn shell. The tool is properly configured, and works in all cases except when using scripting variables.
I have an SQL script, that looks like this.
SELECT COLUMN1 FROM TABLE WHERE COLUMN2 = '$(param1)';
And I'm running the sqlcmd command like this.
sqlcmd -S server -d database -U user -P pass -i input.sql -v param1="DUMMYVALUE"
When I execute the above command, I get the following error.
Sqlcmd: 'param1=DUMMYVALUE': Invalid argument. Enter '-?' for help.
Help lists the below syntax.
[-v var = "value"...]
Am I missing something here?
You don't need to pass variables to sqlcmd. It auto picks from your shell variables:
e.g.
export param1=DUMMYVALUE
sqlcmd -S $host -U $user -P $pwd -d $db -i input.sql
In the RTP version (11.0.1790.0), the -v switch does not appear in the list of parameters when executing sqlcmd -?. Apparently this option isn't supported under the Linux version of the tool.
As far as I can tell, importing parameter values from environment variables doesn't work either.
If you need a workaround, one way would be to concatenate one or more :setvar statements with the text file containing the commands you want to run into a new file, then execute the new file. Based on your example:
echo :setvar param1 DUMMYVALUE > param_input.sql
cat input.sql >> param_input.sql
sqlcmd -S server -d database -U user -P pass -i param_input.sql
You can export the variable in linux. After that you won't need to pass the variable in sqlcmd. However, I did notice you will need to change your sql script and remove the :setvar command if it doesn't have a default value.
export dbName=xyz
sqlcmd -Uusername -Sservername -Ppassword -i script.sql
:setvar dbName --remove this line
USE [$(dbName)]
GO
I think you're just not quoting the input variables correctly. I created this bash script...
#!/bin/bash
# Create a sql file with a parameterized test script
echo "
set nocount on
select k = '-db', v = '\$(db)' union all
select k = '-schema', v = '\$(schema)' union all
select '-', 'static'
go" > ./test.sql
# capture input variables
DB=$1
SCHEMA="${2:-dbo}"
# Exec sqlcmd
sqlcmd -S 'localhost\lemur' -E -i ./test.sql -v "db=${DB}" -v "schema=${SCHEMA}"
... and tested it like so:
$ ./test.sh master
k v
------- ------
-db master
-schema dbo
- static
I am trying to write a shell script to check database connectivity. Within my script I am using the command
sqlplus uid/pwd#database-schemaname
to connect to my Oracle database.
Now I want to save the output generated by this command (before it drops to SQL prompt) in a temp file and then grep / find the string "Connected to" from that file to see if the connectivity is fine or not.
Can anyone please help me to catch the output and get out of that prompt and test whether connectivity is fine?
Use a script like this:
#!/bin/sh
echo "exit" | sqlplus -L uid/pwd#dbname | grep Connected > /dev/null
if [ $? -eq 0 ]
then
echo "OK"
else
echo "NOT OK"
fi
echo "exit" assures that your program exits immediately (this gets piped to sqlplus).
-L assures that sqlplus won't ask for password if credentials are not ok (which would make it get stuck as well).
(> /dev/null just hides output from grep, which we don't need because the results are accessed via $? in this case)
You can avoid the SQL prompt by doing:
sqlplus uid/pwd#database-schemaname < /dev/null
SqlPlus exits immediately.
Now just grep the output of the above as:
if sqlplus uid/pwd#database-schemaname < /dev/null | grep 'Connected to'; then
# have connectivity to Oracle
else
# No connectivity
fi
#! /bin/sh
if echo "exit;" | sqlplus UID/PWD#database-schemaname 2>&1 | grep -q "Connected to"
then echo connected OK
else echo connection FAIL
fi
Not knowing whether the "Connected to" message is put to standard output or standard error, this checks both. "qrep -q" instead of "grep... >/dev/null" assumes Linux.
#!/bin/bash
output=`sqlplus -s "user/pass#POLIGON.TEST " <<EOF
set heading off feedback off verify off
select distinct machine from v\\$session;
exit
EOF
`
echo $output
if [[ $output =~ ERROR ]]; then
echo "ERROR"
else
echo "OK"
fi
Here's a good option which does not expose the password on the command line
#!/bin/bash
CONNECT_STRING=<USERNAME>/<PASS>#<SID>
sqlplus -s -L /NOLOG <<EOF
whenever sqlerror exit 1
whenever oserror exit 1
CONNECT $CONNECT_STRING
exit
EOF
SQLPLUS_RC=$?
echo "RC=$SQLPLUS_RC"
[ $SQLPLUS_RC -eq 0 ] && echo "Connected successfully"
[ $SQLPLUS_RC -ne 0 ] && echo "Failed to connect"
exit SQLPLUS_RC
none of the proposed solutions works for me, as my script is executed in machines running several countries, with different locales, I can't simply check for one String simply because this string in the other machine is translated to a different language. As a solution I'm using SQLcl
https://www.oracle.com/database/technologies/appdev/sqlcl.html
which is compatible with all sql*plus scripts and allow you to test the database connectivity like this:
echo "disconnect" | sql -L $DB_CONNECTION_STRING > /dev/null || fail "cannot check connectivity with the database, check your settings"
#!/bin/sh
echo "exit" | sqlplus -S -L uid/pwd#dbname
if [ $? -eq 0 ]
then
echo "OK"
else
echo "NOT OK"
fi
For connection validation -S would be sufficient.
The "silent" mode doesn't prevent terminal output. All it does is:
-S Sets silent mode which suppresses the display of
the SQL*Plus banner, prompts, and echoing of
commands.
If you want to suppress all terminal output, then you'll need to do something like:
sqlplus ... > /dev/null 2>&1
This was my one-liner for docker container to wait until DB is ready:
until sqlplus -s sys/Oracle18#oracledbxe/XE as sysdba <<< "SELECT 13376411 FROM DUAL; exit;" | grep "13376411"; do echo "Could not connect to oracle... sleep for a while"; sleep 3; done
And the same in multiple lines:
until sqlplus -s sys/Oracle18#oracledbxe/XE as sysdba <<< "SELECT 13376411 FROM DUAL; exit;" | grep "13376411";
do
echo "Could not connect to oracle... sleep for a while";
sleep 3;
done
So it basically does select with magic number and checks that correct number was actually returned.