Well, if you use 'anything' for several databases, it will be a single point of failure (SPOF). In a RAC env, you will have an ASM instance on each node, so there is no SPOF (beside your disks, but they are usually mirrored, so the risk is lessened). Last, high availability is not an illusion with RAC, I've got 2 clients each using a 2-nodes RAC, and each one experienced a system failure on one of the nodes, and noticed it 24 hours later, all the users and apps were using the other instance and nobody saw that one machine was down until they noticed a small increase in response times, looked for the cause and found out that one instance was down. Granted, there are bugs in all Oracle's versions, but the same is true for all the vendors.
rgds
On 05/07/2006 02:53:16 PM, Ranko Mosic wrote: > Hi List, > Just thinking aloud - when ASM crashes then all instances connected to it > crash too. > This can't be good, right ?
Well, it would have been a single point of failure if Oracle wasn't unbreakable. ASM is, of course, bug free, so it will never crash. What is even more important, it makes it possible to forgo buying the expensive clustering software from SUN, HP or IBM and save some money, while providing an illusion of high availability.
-- Mladen Gogala http://www.mgogala.com
-- //www.freelists.org/webpage/oracle-l
