Data ONTAP provides database file protection through the nvfail option of the vol options command. The nvfail option enables a filer to detect NVRAM inconsistencies at boot time. Use it to warn database administrators of NVRAM problems that could compromise database validity. If the filer finds any problems, database instances hang or shut down, and the filer sends error messages to the console to alert you to check the state of the database.
You can provide additional protection to specific database files by adding them to an optional file you create called /etc/nvfail_rename. When you enable nvfail and the filer detects NVRAM errors at boot up, the filer renames any database files specified in the nvfail_rename file by appending .nvfail to the original file name. When the filer renames a database file, the database cannot restart automatically. This gives you the opportunity to examine the file for inconsistencies before you remove the .nvfail extension and make the file accessible again.
When you enable nvfail, the following process takes place during boot-up.
See the documentation for your specific database software for instructions about examining database file validity.
When you enable nvfail and the filer encounters NVRAM errors, the message sent to the console and the /etc/messages file looks like the following:
All filehandles have been invalidated due to previous NVRAM failure.
All filesystems must be remounted by the client(s).
If you created an nvfail_rename file, you receive an additional message:
Renaming files in /etc/nvfail_rename.
nvfail_rename: old_file_name: new_file_name
Copyright © 1999, Network Appliance, Inc. All rights reserved.
Part No. 210-02334
File: net24.htm
Last updated: 04/26/99 08:11:18