Shutdown immediate hangs when used in backup scripts
Posted on 2005-05-09
Occasionally our weekly cold backup script fails due to receiving a 'shutdown: waiting for active calls to complete' message in the database's alert log (this situation occurs across a number of databases at various times, never any one specific database). I've researched this issue here, at Metalink and around the internet in general but have not seen a definative answer to the problem. Ask Tom website gives as resolution of "do not perform cold backups" which does not seem like an appropriate answer. I've heard that an approach to this problem might be to add code into the existing backup script to locate and 'kill' any client connections to the database(s), specifically looking for SID processes that contain (LOCAL=NO). Is there a better approach to this problem (other than scripting) and if not, is there anyone here that might have a script (or portions of their backup script) that would enable such process checking/termination as part of their current weekly cold backup script? Our platform is Solaris 2.6/2.8 and Oracle version 220.127.116.11/18.104.22.168-6 and 10.1.0.4 (note: no such shutdown issues have been noted in 10.1.0.4). This problem has been a long time issue, it would be great to finally work around it (or even better, resolve it). Thanks for your assistance.