We have our PROD database running with 2 different
archival destinations. One of the arch files goes to
the local disks and the other one to NFS file system.
Sometimes, the NFS file system goes down for
a couple of minutes and so the ARCH process stops
writing to that destination. The ARCH process does
not write to this destination again.
The only thing we could currently do is to stop and
restart the Oracle server in which case, the ARCH starts
writing to both the file systems again.
Is there any way we can restart archiving to the second
destination without shutting down the server.
Thanks for your help,
Change your dest. to local disk (2 desti) because once one of desti. fails, they all failed. you have to re-start ur oracle to make it re-work again.
you can have 2 desti. all in local disk(s) and periodically copy old archived files to your NFS storage.
I don't know if this help.
If I understand your question, you can use:
alter system archivelog start;
How do you switch between the destinations ?
Actually, having a NFS mount as arch log destination may not be a good idea as things will most likely get slowed down.
Why cann't you stick only to the local disk and have a daemon process copy you archive logs from this disk to the NFS mount ? That way, you can always keep on writing to local disk and still keep its utilization under control.
And I think the archival process starts automatically once the destination is fine. It doesn't call for restart of the database.
May be that you have some different problem.
Try going thru this parameter setting in your initilization file of your database
The above parameters default value is 1 if set will continue to write to log_archive_dest_1 and log_archive_dest_2. I would consider ur NFS to be on log_archive_dest_2
So if the above parameter is set to 1 it has to write the archive files to log_archive_dest_1 but optional for log_archive_dest_2.If the log_archive_min_suceed_dest parameter is set to 2 it has to compulsorily write to both the destination which I would not suggest because you said ur NFS is not reliable.
so you are database server continues to run and whenever the NFS daemon should come up it should start writing the archive log files back to NFS
Try this on a test system if it suceeds you can implement on ur production.
Hope this helps
Click Here to Expand Forum to Full Width