Backup in NT and Linux

[amolso]

Hello

I have made the following setup in my NT init.ora file
however the last log archive format is commented.

Do I have to uncomment It ?

Initially log_archive_start was false and no destination directories. These are added by me.
I have to ask after setting database into archivemode
and enabling the automatic archival.

log_archive_start = true
log_archive_dest_1 = "location=D:\Oracle\oradata\mdb\archive"
log_archive_dest_2 = "location=G:\oradata\marchive"
log_archive_dest_3 = "location=H:\oradata\marchive"
# log_archive_format = %%ORACLE_SID%%T%TS%S.ARC

I my Linux machine with 8.1.5 in init file there is nothing mentioned about 'log archive start ' parameter.
Do I have to write code above in the similar manner in linux oracle parameter as well

Amol New to dba world, so please ......

 

[karluk]

Archiving is pretty much the same on all operating systems. It's part of the Oracle architecture, so what you have to do on Linux is the same as on NT.

1. The log_archive_format is not essential. This parameter determines the names of the archive log files. If you don't have log_archive_format in your initialization file (or have it commented out), Oracle will use a default for the file names.

2. Log_archive_start and at least one log_archive_dest must be supplied.

3. There is one other thing that you have to do, in addition to setting these initialization parameters. Mount the database, but do not open it. Then enter the command "alter database archivelog;" Then open the database and run the command "archive log list" to verify that the database is running in archive log mode and automatic archiving is enabled.

 

[amolso]

Thanks Karluk

I have another query based on above
Currently I have 3 destinations supplied for backup as above in my Windows NT Oracle Machine. When Explore thru the files I found that the sizes of files are very much different. Currently I am just running a trial database where very few data entry activity is going on and my archivemode and automatic archival are properly set. In my all three destinations the files are same with same size. But the size of Arc00304.001 is much much less than Arc00306.001 which is created next day only while file Arc00305.001 is less than first which was also on same day. Files created on next few days by the system are also varying in the size. Can Anyone please explain what is the status there in these archive files ? Are these files are backup of only control files ? DO these Files I have to keep or old files has to be deleted?
Does these file backs up only data currently entered in the day or the old data as well in the tables ?
Which of these files will be required to store on the outside disks in case of disk or system crash. There are also 2 winzip files with .ARC extensions are created for last two days Do I need to back them up as well ? What does these winzip file contain ?
Do I have to run the tablespace backup every day or any specific interval for datafiles ?

Thanks in Advance
New to dba world, so please ......

 

[karluk]

I will do my best to answer your questions. However I think you should also invest in a book on Oracle backup and recovery. I'm sure that any decent book on the subject would cover things more thoroughly than we will in this forum.

1. The size of your archive logs should generally be the same size as your redo logs. They should be exactly the same size unless a log switch happened before the current redo log filled up.

Please note that your redo logs don't have to be the same size. Usually all the redo log groups are created with the same size, but I'm fairly sure that this isn't a requirement. That is one possible reason that you are seeing different sizes on your archive files.

2. Your archive logs are most definitely not a backup of your control files. The control files need to be backed up separately. Failure to do so could easily get you into an unrecoverable situation. There are two ways to back up control files, both of which can be very useful. "alter database backup controlfile to trace" generates a sql statement that can be used to recreate the control file from scratch. "alter database backup controlfile to 'full_path_name'" makes an exact copy of the control file in the specified location.

3. Archive logs contain database update activity for a certain time period. That means you need all the archive logs that were generated since the last time the database data files were backed up. After that they can be deleted.

4. Any good backup strategy will include copying essential files to tape or some other storage medium. Then you won't have to worry about a disk crash destroying your database. You should also consider storing backups at a completely different location in another building. Then you won't be as vulnerable to fire or some other natural disaster.

5. Database recovery should be tested long before you have a failure. The worst time to find out that you are missing some essential file is when Oracle is telling you that it is needed for recovery. Any good DBA will make time for testing, and I think it's especially important in your situation, considering your relative inexperience.