| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Usenet -> c.d.o.misc -> Re: [q] How to pick up archived logs
On 20 Feb 1997 08:19:05 +0300, "Vasily R. Kraskovsky" <vasily_at_kvr.spb.ru> wrote:
>Hi!
>
> I have Oracle 7.3.2 under Solaris 2.5 on my place, and I need to switch
>database backup mode from full-offline backup to online backup, trying to
>attain 24-hour database accsessibility. So I want to pick up archive logs
>when they are already finished archiving, move them to another server, and
>there add/compress them to solid archive.
>
> Question: How can I detect from unix (ksh script starting by cron) that
>Oracle have completed another log to archiving, and it can be moved to
>another place? Help, please.
>
>Thanks.
>
>--
> Nemo solus sapit satis
> <<<<<<<<<<<<<<<<<<<<<<< >>>>>>>>>>>>>>>>>>>>>>>
> Kraskovsky Vasily, DBA, Petersburg-Express, SPb, Russia
> vasily_at_kvr.spb.ru
>
>
Here is the cron script I use on my system to compress and copy my archived redo log files:
#!/bin/csh -f
setenv ORACLE_HOME /usr/oracle
setenv ORACLE_SID oracle
set path=($path $ORACLE_HOME/bin /opt/gnu/bin)
cd /d01/arch
echo set heading off > /tmp/$$.sql
echo set feedback off >> /tmp/$$.sql
echo set linesize 80 >> /tmp/$$.sql
echo column cmd format a70 >> /tmp/$$.sql
echo spool /tmp/$$.sh >> /tmp/$$.sql
echo select "'gzip ' || archive_name || " >> /tmp/$$.sql
echo "'; cp ' || archive_name || '.gz /d05/arch' cmd">> /tmp/$$.sql
echo 'from v$log_history' >> /tmp/$$.sql
echo 'where archive_name in(' >> /tmp/$$.sql
ls /d01/arch/*.log | sed "s/^/'/;s/"'$'"/',/" >> /tmp/$$.sql
echo 'NULL ) and sequence# not in ' >> /tmp/$$.sql echo '( select sequence# from v$log where archived = '"'NO'"')' >> /tmp/$$.sql echo '/' >> /tmp/$$.sql
sqlplus tkyte/tkyte @/tmp/$$.sql
sh /tmp/$$.sh
rm /tmp/$$.{sh,sql}
--------------------------- eof ------------------------------------
basically I'm building a query in /tmp/$$.sql that will (based on the files existing in my archive dest when the script begins) select out a 'gzip filename; cp file somewhere_else' command from the database for those logs that have been archived completey (i use the file system to tell me which files I already copied, anything that has an extension of .gz won't get processed twice). The query runs and builds another sh script which does the actual gzipping and copying of files.
Thomas Kyte
Oracle Government
tkyte_at_us.oracle.com
 |
 |