Printable View
Hi,
What will happen if the archive log distination is full and redo files can't be archived?
Will the instance hang? And automatic archive will stop by itself?
How to solve this without shuting down the instance?
And how can we avoid this situation? Is there any alternative destination that can be specified in the INIT file so that if the primary destination is full , ORACLE starts archiving at secondary destination without any problem.
Could any one explain my queries?
If the direectory for archiving is full you will see lot of erros in alert.log saying that. You will free space in that directory and the database will continue normally.
If all your redo ogs are full and the directory for archiving is full, then the database will stopped, waiting for you to free space. sometimes I saw a database go down when the archiving is not possible and all your redo logs are full.
In relation to your last question, I think it is not possible to do that.
Hope this will help to solve your doubts
Angel
I'm not aware of any init.ora parameters to achieve what you want. However you could write a script to take care of this for you using the shell, perl etc. and schedule it to be run every x minutes by the cron/NT at jobs etc. (sorry don't have one myself otherwise I'd share it). One thing to bear in mind is not to touch the most recent archive log file in your script, as it may be getting written to as your script executes.
What I suggest is RMAN and the command CONFIGURE. No need to write any scripts. What you suggest was done before RMAN. Oracle has developed enough, so writing OS scripts is very little needed in nowadays Oracle.
See:
http://download-west.oracle.com/otnd...dbylogappx.htm
Log_archive_dest_n (n=1...9), but I think I had some problems in 9i with this, I'm not very clear why... The known problem of "option not installed: managed standby" - known, but not by me :) - , I had problems when I set it on installation point. I think is better if you make a db with the wizard to set it norchivelog at first, then change the state. But I'm not sure why, maybe I'm wrong, or it is a bug, or even Oracle changed something without telling...
With this parameters you will have exactly copies of archives, but if the first stops, all are stopped.Quote:
Originally posted by ovidius
See:
http://download-west.oracle.com/otnd...dbylogappx.htm
Log_archive_dest_n (n=1...9), but I think I had some problems in 9i with this, I'm not very clear why... The known problem of "option not installed: managed standby" - known, but not by me :) - , I had problems when I set it on installation point. I think is better if you make a db with the wizard to set it norchivelog at first, then change the state. But I'm not sure why, maybe I'm wrong, or it is a bug, or even Oracle changed something without telling...
Cheers
Angel
Oh, that means I understood totally wrong way this problem... :) Sorry!
I have somehow another problem: keeping log_archive_dest "clean" is simple. The problem is if I move the archives on different tapes, how will I proceed in case of a disaster recovery? What will be the "unbreakable" :D methodology in this case?
But again, I did have a reason to understand it is possible to give an alternative (from that link):
"Use the ALTERNATE attribute of the LOG_ARCHIVE_DEST_n parameters to define an alternate archiving destination to be used if archiving to the original archiving destination fails.
"
Maybe you have reason and I talked very quickly ;)Quote:
Originally posted by ovidius
But again, I did have a reason to understand it is possible to give an alternative (from that link):
"Use the ALTERNATE attribute of the LOG_ARCHIVE_DEST_n parameters to define an alternate archiving destination to be used if archiving to the original archiving destination fails.
"
Cheers
Angel
Hi,
If my archivelog directory/filesystem is full, I'll move old archivelogs to free up space and then attempt to 'alter system switch logfile'.
This doesn't seem to work, even though there is now space on the filesystem to write the archivelog to disk, a shutdown and restart of the DB are required.
Is there any way to archive the log after freeing up space without restarting the instance?