This post is based on an email i’ve answered to a person from the Solaris Users list, so it is written in that context.

If you are using cpio to backup, you may want to check with the script the value returned by the backup operation and based on the returned value, send an email or not, you would have to check which error code the cpio returns on a failure, you can verify the $? variable (dollar and question mark) after the cpio command line inside the script… but keep reading.

A second way would be to check the logs while the backup is running and to use a pattern on the raw text to find errors and then to take the decision…keep reading.

The third (and the best) is to use the solaris tool fssnap (to take a snapshot of the filesystem), this tool is used to backup live environments like your database; you still have to stop the database but you only stop it while the snapshot is being taken, which is much faster than doing the backup, when the filesytem snapshot has been taken you start the database again and then you can take your time to do the backup smootly…

To finish, you could also send an SMS to your cellphone or a beeper reporting the error if it is mission critical.