|
I don't recall exactly how I came about the knowledge, but I was
surprised to find out how much stuff that's very specific to ASM
happens in the dedicated server process and that's why I'm thinking
this may at least be explainable. I still don't necessarily agree that
it *should* be that way, but maybe we can at least learn if that's the
reason for this symptom. If no one else can test it, I should be able to test it out next week. Dan Bobak, Mark wrote: -- //www.freelists.org/webpage/oracle-lWell, it's just a theory. But I don't see why ASM traffic would have anything to do with the interconnect...? I'm not currently able to test the theory, but maybe someone else can. Personally, I'd expect a single instane db to be able to detect and connect to the local node of a clustered ASM installation....but maybe not.... I just don't know enough about the internal workings of ASM... -Mark -- Mark J. Bobak Senior Database Administrator, System & Product Technologies ProQuest 789 E. Eisenhower, Parkway, P.O. Box 1346 Ann Arbor MI 48106-1346 +1.734.997.4059 or +1.800.521.0600 x 4059 mark.bobak@xxxxxxxxxxxx www.proquest.com www.csa.com ProQuest...Start here. -----Original Message----- From: Dan Norris [mailto:dannorris@xxxxxxxxxxxxx] Sent: Friday, April 04, 2008 1:53 PM To: Bobak, Mark Cc: Jeffery Thomas; oracle-l Subject: Re: Cluster Interconnects configured in single-instance DB Mark, I'm not sure I agree with that theory. When an instance uses ASM, part of the ASM codepath is executed by the server process when accessing ASM. So, it makes some sense to me that the RDBMS server would have some potential need to access the interconnect when using ASM. I tend to agree with your (Jeff) theory that it's got cluster interconnects defined because of the clustered ASM interaction. That makes more sense to me than because of the binaries it uses. I suppose the test would be to install a new O_H, do "make -f ins_rdbms.mk rac_off", then start up this DB using those binaries to see what happens. My bet is that if you disable the RAC option and have clustered ASM, you won't be able to start the instance, but I don't have a suitable test environment to find out for sure. Hopefully someone does or I'll have to build one and find out because now I have to know the answer :). Dan Bobak, Mark wrote:Well, it can't be really *using* it, in the sense that it's got no other nodes to talk to, but perhaps the RAC-aware binaries have a code path that detects the interconnects (apparently by reading the OCR?) and lists them in that V$ view? -----Original Message----- From: Jeffery Thomas [mailto:jeffthomas24@xxxxxxxxx] Sent: Friday, April 04, 2008 1:27 PM Subject: Re: Cluster Interconnects configured in single-instance DB Yes it does -- and in fact, I thought of that 5 seconds after I sent this email. I guess then, it is actually *using* it? On Fri, Apr 4, 2008 at 1:18 PM, Bobak, Mark <Mark.Bobak@xxxxxxxxxxxx> wrote:Hmm....does the single instance database use the same $ORACLE_HOME as the RAC database? -----Original Message----- From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Jeffery Thomas Sent: Friday, April 04, 2008 12:56 PM To: oracle-l Subject: Cluster Interconnects configured in single-instance DB We have a two-node 10.2.0.3 RAC system on Solaris 10 using Oracle Clusterware and ASM that is being used for our test database.. We also have a single-instance database on one node of this cluster being used for development. The dev database is using ASM, sharing the same data disk group as the RAC database. I noticed on startup in the alert.log that it is configuring interconnects; in addition, when I perform this query in the single-instance database: SQL> select * from v$cluster_interconnects; NAME IP_ADDRESS IS_ SOURCE --------------- ---------------- --- ------------------------------- e1000g2 192.168.1.103 NO Oracle Cluster Repository e1000g5 192.168.1.102 NO Oracle Cluster Repository The question I have is: why? Is it because it is accessing a disk group maintained by a clustered ASM? Thanks, Jeff |
