Ever wonder why your Oracle DBA doesn’t do much on Friday? Well they are trying their darnest not to break anything.
I have had several examples of Oracle Software BREAKING!…Crunch there it goes. I am talking repeatable, consistent but intermittent events.
1. Grid Control Oracle Management Server 10.2.x core dumping every few seconds filling up the hard drive completely. (This has happened to several 10.x installs) No ideas what causes it because it is impossible to troubleshoot with a full drive. Reinstalling the OMS (using the same repository) is the easiest fix I came up with.
2. Cloned Oracle Application Server (OAS) Installs – Configuration working fine for months serving up Oracle Forms. All of a sudden complete loss of service, rebooting doesn’t fix the problem. We run multiple OAS behind a load balancer. taking one down or removing it from the balanced pool is easy. I now do a manual install of Oracle Application Server.
3. Oracle Application Server going bad – This is a manually-installed version that gets corrupted somehow. Won’t restart, rebooting doesn’t fix things. Only think I can do is restore…you do know there is an OAS backup and restore utility? It works well but takes up quite a bit of room, I backup to a local drive whenever there is a major change in the OAS configuration that needs to be saved. Again this OAS can be taken out the load balanced pool of servers. This has happened on a intermitten basis on both Solaris and Red Hat Linux.
These are items that I don’t enter an Oracle Service Request for because that takes more TIME than doing a system-type restore. Recently applied a PSU to a Grid Control 10.2.0.5 install which completely buggered the system, all Oracle had me do was install, uninstall umpteenth times. Exact same results all times, I demonstrated the install procedure with the engineer. I closed this contact due to lack of resolution. After waiting a couple of weeks, I restored the system so I could start using GC again.
Anyone else have an example they want to share?