Re: I need one opinion and one answer

Raid5 should protect you from a data loss in the event of 1 disk failing. It won't protect you if you do something like :

   $ rm database_file

You can recover from media failure if you have:

1. Lost archive logs, but have all the database files and online logs intact or
2. Lost some or all database files but have backups of them and ALL archivelogs since the backup was done. If you lose 1 archive log, you can only rollforward to that point.

If you lose the current online log, you will lose some data and will need to recover all files from backup and rollforward. You don't want this to happen. This is why we mirror the online logs.

Also, with raid, it stripes the data across all the drives in the raid set, so it doesn't really matter what file is on what disk. Raid5 also gives a significant write performance penalty, but reads perform well.

Regards,
Andrew.

Van Messner wrote:
>
> Opinions Please
>
> I'm monitoring a group of databases running on Novell servers. Each
> server has two disks (each with its own controller) which hold mirrored
> control files and mirrored online redo logs. Everything else goes on the
> server's raid5 setup - including the archived redo logs. This means I have
> no control over which disks hold data, indexes, archived logs etc.
> Normally I'd want to see the archived redo logs separate from the data
> files (and even mirrored), but I don't know enough about raid5 to tell
> whether we're protected against disk failure. The company takes a complete
> cold backup of each database every night. But they say they can't afford to
> lose part of a day's data if we crash after some of the redo logs have been
> archived. Is this a danger or does raid5 protect us?
>
> Question
>
> For each database the init.ora file sets archiving on and sets a
> destination for the archived redo logs. I can look at the alert log and see
> that log switches take place with some regularity on each database. When I
> do a backup controlfile to trace I see that nine of the databases have
> archivemode set to yes and one has it set to no. The one that has it set to
> no is not producing any archived redo logs. Since each database is shut
> down, backed up and started again the following morning, and since each
> init.ora sets archiving on, how or why does the one database not archive the
> logs? Where is the setting being overridden?
>
> Thanks,
> Van
Received on Wed Feb 10 1999 - 23:10:06 CST