First off, Oracle 8.1.7.4 on Solaris 8...

During a moderately heavy load, the database sent the "Failed to archive log" message to the alert log. I understand that this is an informational message and can be ignored. But the issue is that around the time that this message appeared in the alert log, the database seemed to hang. Logins would take up to 30 seconds and active sessions would stall. I think the problem is with I/O contention with the redo logs and the archive log processes as they are on the same physical drives. The Log File Sync wait event is relatively high compared to the other wait events ( ~ 400ms/sec peak ).

Can anyone confirm my theory, offer an alternative reason for this behavior or point me to some other reasons? Thanks,

John