Comments on: Oracle databases with ASM in normal redundancy mode http://blog.ronnyegner-consulting.de/2009/12/04/oracle-databases-with-asm-in-normal-redundancy-mody/ Ronny Egners Blog about Oracle, UNIX and EMC / Legato Networker Mon, 23 Aug 2010 06:12:32 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Ronny Egner http://blog.ronnyegner-consulting.de/2009/12/04/oracle-databases-with-asm-in-normal-redundancy-mody/comment-page-1/#comment-561 Ronny Egner Sat, 19 Dec 2009 09:05:59 +0000 http://blog.ronnyegner-consulting.de/?p=1419#comment-561 Hi, i certainly agree with you. The third voting disk was located on a third site accessible via NFS. With the same system Oracle 11g R1 was far more stable than 10.2.0.4.2 (with installed PSU 2). I had nothad the time to investigate this further. The question here is: Did you really tested it? It SHOULD be not problem loosing one failure group. But in fact 10g R2 crashes. Ronny Hi,

i certainly agree with you. The third voting disk was located on a third site accessible via NFS. With the same system Oracle 11g R1 was far more stable than 10.2.0.4.2 (with installed PSU 2). I had nothad the time to investigate this further. The question here is: Did you really tested it? It SHOULD be not problem loosing one failure group. But in fact 10g R2 crashes.

Ronny

]]> By: Martin Decker http://blog.ronnyegner-consulting.de/2009/12/04/oracle-databases-with-asm-in-normal-redundancy-mody/comment-page-1/#comment-543 Martin Decker Fri, 18 Dec 2009 08:46:32 +0000 http://blog.ronnyegner-consulting.de/?p=1419#comment-543 Hello, I have performed several stability tests on different customer sites with CRS/ASM/RDBMS 10gR2 (10.2.0.4) and normal redundancy. During these tests, the impact of loosing one of two storages on availability of server and database should be minimal, certainly no crashes or even unstartable instances. What´s important in case of two storages is the location of the third voting disk because of the majority rule. However, the ASM diskgroup availability will not be impacted by loosing one failgroup. Regards, Martin Hello,

I have performed several stability tests on different customer sites with CRS/ASM/RDBMS 10gR2 (10.2.0.4) and normal redundancy. During these tests, the impact of loosing one of two storages on availability of server and database should be minimal, certainly no crashes or even unstartable instances.

What´s important in case of two storages is the location of the third voting disk because of the majority rule.

However, the ASM diskgroup availability will not be impacted by loosing one failgroup.

Regards,
Martin

]]>