Solved

Oracle RAC ASM setup on a shared Netapp SAN partition in RHES

Posted on 2008-06-16
2
2,469 Views
Last Modified: 2013-12-15
Hi,

We have 3 RHES servers (V4.6) - the DBA would like Oracle RAC ASM on all 3 servers
access a common NetApp SAN raw disk partition.

We found that /dev/mapper/mpath1 on  server 1  is /dev/mapper/mpath0  on  server 2
& yet a different device name on the 3rd server.  Good thing is /dev/mapper/mpathx
stays permanently linked to the same disk partition each time the RHES servers are
rebooted.  If we use /dev/mpath/pathx, it changes each time servers are rebooted,
ie on server 1, after a reboot, the value in  /dev/mpath/pathx  will need to be changed
so as to access back the same disk partition but if we used /dev/mapper/pathx, it's
stays the same after reboots.

The DBA has set up Oracle RAC clustering (using sort of ASM tool - I'm not sure as
I'm no DBA but Unix Admin).  She's sure that on all the servers, she has used the
same SAN disk partition & is able to create the database successfully.  So on all
the servers,  ORACLE_SID=+ASM2  & when she went into each of the servers to
do a "select" to retrieve data, only on server1 it returns data while on the other
servers, nothing was returned.

I'm not sure how the DBA ensure she's accessing the same partition - is there anyway
I can verify this other than
" formatting (using fdisk) on each of the server the disk partition
to a  character device, mount the formatted character device to a mount point on each
server, create a dummy file, say test.txt & go into the other 2 servers to check if test.txt
is there " - to do this, she'll lose her work done so far.

Also, if RAC clustering is used, is Linux clustering still needed or RAC clustering can
coordinate the IO between the servers?  If Linux clustering is needed, kindly let me
know the Linux command to start up the clustering service

We have another set of older Linux servers which the outsourced DBA told us
they used /dev/raw...   device files which also connects to Netapp SAN disk partitions
& no Linux clustering is running but this set of servers is owned by the outsourced
DBAs/ vendors & I have yet o verify what she said is true.

Lastly, we're contemplating if we should use character device instead of raw device.
Can Oracle ASM be used on a non-raw (character) device Unix file system?

I'm completely newbie to Oracle/RAC/ASM & NetApp SAN, so bear with me.
0
Comment
Question by:sunhux
[X]
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
2 Comments
 
LVL 48

Assisted Solution

by:schwertner
schwertner earned 150 total points
ID: 21800693
RAC clustering uses so called "Cash Fussion".
This means that the caches (SGA) of all nodes
establish a common area accessable from
all instances on the RAC. So if a table row is read by one
node all other nodes will use it without reading it physically from the SAN.
0
 
LVL 16

Accepted Solution

by:
Richard Olutola earned 350 total points
ID: 21801507
>Also, if RAC clustering is used, is Linux clustering still needed or RAC clustering can
coordinate the IO between the servers?  If Linux clustering is needed, kindly let me
know the Linux command to start up the clustering service

NO. If you're implementing RAC you DO NOT need any other O/S clustering.

>Lastly, we're contemplating if we should use character device instead of raw device.
Can Oracle ASM be used on a non-raw (character) device Unix file system?

You use either Filesystem (LVM, ocfs2...), raw device, or ASM. If you are planning to use ASM, then you cannot use raw device together. ASM simplifies file management and provides DR and performance enhancements.

Best performance can be achieved with raw devices but it's a pain to manage. For instance, a raw partition is required for each datafile!

>The DBA has set up Oracle RAC clustering (using sort of ASM tool - I'm not sure as
I'm no DBA but Unix Admin).  She's sure that on all the servers, she has used the
same SAN disk partition & is able to create the database successfully.  So on all
the servers,  ORACLE_SID=+ASM2  & when she went into each of the servers to
do a "select" to retrieve data, only on server1 it returns data while on the other
servers, nothing was returned.

Note that the SID is different on each node. If SID is +ASM1 on node1 then it would be +ASM2 on node2 etc.

R.


0

Featured Post

VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

The question appears often enough, how do I transfer my data from my old server to the new server while preserving file shares, share permissions, and NTFS permisions.  Here are my tips for handling such a transfer.
A look at what happened in the Verizon cloud breach.
Via a live example, show how to take different types of Oracle backups using RMAN.
This video shows how to copy an entire tablespace from one database to another database using Transportable Tablespace functionality.
Suggested Courses

617 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question