I have always thought the consequences of changing CPU dynamically, such as HP Superdome Virtual Partitions. There are quite a few oracle parameters which depend on cpu_count, what happens to them if we change the number of CPUs all day? On 2/17/06, Scott Canaan <srcdco@xxxxxxx> wrote: > > We are in the process of implementing Sun Solaris 10 with zones and > Oracle 10.2.0.1. As DBA's, we were told that this was what the > environment was going to be, without any request for feedback or testing. > The Oracle home is located in the global zone and each Oracle instance is in > its own local zone. We've had to allow the local zones permission to write > to the global zone because the Oracle database and listener wouldn't even > run if we didn't. According to the SA, this violates the security that > zones was supposed to give us, but he reluctantly did it. > > This week, he's been playing with dynamically changing the number of > processors and the amount of memory allocated to the zones. This has caused > all kinds of problems with Oracle. We've had several crashes that have led > us to add the init.ora parameter _enable_numa_optimization=false to all of > Oracle databases running in zones. According to metalink, this is a bug > that is supposed to be fixed in 10g. Last night, we had a database crash > because the memory changed dynamically and it wasn't happy about it. > > My question is: Is anyone else running Oracle in Solaris 10 zones? If > so, what have you had to do to keep the instances up and running? What > problems have you encountered along the way? > > > > Thank you, > > > > Scott Canaan '88 (Scott.Canaan@xxxxxxx) > > (585) 475-7886 > > "Life is like a sewer, what you get out of it depends on what you put into > it." - Tom Lehrer. > > >