Solved

detail information about Oracle Stretched (Extended Distance) Clusters

Posted on 2014-09-04
11
374 Views
Last Modified: 2014-09-29
0
Comment
Question by:marrowyung
  • 7
  • 3
11 Comments
 
LVL 1

Author Comment

by:marrowyung
ID: 40302975
I want to focus on
1) how failover to another local node using RAC
2) how to failover to another remote nodes using RAC over a WAN.
3) how to apply patches for each node of the RAC.
0
 
LVL 28

Expert Comment

by:Ryan McCauley
ID: 40325797
Are you looking for an actual walk-through of setting up RAC? There are a number of papers that talk about the limitations and considerations when choosing to deploy RAC over an extended distance, but everybody seems to point back to Oracle's white paper as the best resource:

http://www.oracle.com/technetwork/products/clustering/overview/extendedracversion11-435972.pdf

That said, I don't believe setting up or executing a chance on an extended RAC cluster is any different than configuring it when the two nodes are next to each other in the physical rack and connected to the same switch:

http://www.oracle.com/technetwork/articles/virtualization/wartak-rac-vm-099826.html#5b1

In that walkthrough, there's even a segment of section 7 that specifically talks about how to configure things if you're doing extended RAC, so you can test the configuration there. If there's something else you're looking for, please let me know what is missing and I'll try to provide some additional detail.
0
 
LVL 1

Author Comment

by:marrowyung
ID: 40327252
ryanmccauley,

tks for that and you bring me a step forward.

by this http://www.oracle.com/technetwork/articles/virtualization/wartak-rac-vm-099826.html#5b1, this is for Oralce on linux configuration. Any step by step configuration for Windows ?

is there any  Oracle VM for Windows for me to build an execise?

"That said, I don't believe setting up or executing a chance on an extended RAC cluster is any different than configuring it when the two nodes are next to each other in the physical rack and connected to the same switch:"

someone told me RAC only for local site, but not DR site.... they ran oracle for more than 10 years.
0
 
LVL 1

Author Comment

by:marrowyung
ID: 40327259
do you kow if the information from the link should work with 11g R2 and 12c, right?

I can't find any video trainning for that from youtube to speed up the learning speed, can you find any like that or any video has RAC training including local RAC and extended RAC ?

the vidoe trainning I found is not good as when demostrating RAC failover, they just turn off the active node and let RAC detect and failover, it is surprise that the RAC failover but that will take more than 1 minutes on demo machine .

also any documentation on
1) how to replace/repair failed  for local RAC and extended RAC?
2) Any patch procedure for local RAC and extended RAC?
0
 
LVL 1

Author Comment

by:marrowyung
ID: 40327272
0
Free Trending Threat Insights Every Day

Enhance your security with threat intelligence from the web. Get trending threat insights on hackers, exploits, and suspicious IP addresses delivered to your inbox with our free Cyber Daily.

 
LVL 28

Expert Comment

by:Ryan McCauley
ID: 40329167
Here's a step-by-step walk-through for a Windows server (they're running VirtualBox, which is free VM software on a Linux host - however, the guest VM where they install RAC is Windows Server 2008):

http://www.oracle-base.com/articles/11g/oracle-db-11gr2-rac-installation-on-windows-2008-using-virtualbox.php

That would actually go start-to-finish using a VM, so you could test the entire process on your own development workstation and know how it's going to behave. I don't know of any video walk-throughs, but the step-by-step I've provided should cover everything you need to know.

As for the distance of RAC, I've heard as well that RAC isn't feasible over more than about 100km because of the latency you experience - because both servers can edit the database (multiple active front ends, a hallmark of RAC), they need to be in constant communication about open transactions and their data modification intent. It can be set to be tolerant of longer distances and higher latency, but transactional throughput begins to suffer pretty dramatically once you get further apart than that. Personally, I've never implemented RAC beyond two servers in different racks of a single data center - not for fear of the latency, but just because we didn't have the need.

RAC is generally implemented in cases where you need the throughput of multiple active front end servers - where the processing and/or storage bandwidth of a single server isn't sufficient for your needs. If you're really looking for a DR solution, why not consider replicating to a second site using Dataguard? Or setting up a failover cluster instance (where Oracle is running on a single server, but can move to a second server if the first goes offline) with the nodes in different data centers? RAC solves a specific problem and is great at it, but this may not be a case where it's necessary to meet your needs, and the administrative overheard isn't negligible.
0
 
LVL 1

Author Comment

by:marrowyung
ID: 40329650
"If you're really looking for a DR solution, why not consider replicating to a second site using Dataguard? "

we will have it but for DR purpose.

"Or setting up a failover cluster instance (where Oracle is running on a single server, but can move to a second server if the first goes offline) with the nodes in different data centers?"

you mean by active data guard ?

"RAC solves a specific problem and is great at it, but this may not be a case where it's necessary to meet your needs,"

a lot of company here like that as thye can explain to the management that oracle never down. so they prefer to setup a cross site RAC to complete all HA and uptime request/ SLA., no one will kill them then.

But as far as I know the failover of RAC nodes of the same instance can take a long time as cluster failover usually take a long time (2 minutes by some web demo), is that correct?

how long in your case ?

"
 As for the distance of RAC, I've heard as well that RAC isn't feasible over more than about 100km because of the latency you experience "

any best practice on this  ?
 tks.

also the link you show me is setting up VM box under linux and install Windows 2008 on that VM box but not the oracle is setup on Windows already.  or it just show me how to make RAC works on that on that windows VM which have oracle RAC installed on top of Windows ?


btw, how long it takes for the active data guard to failover and switch back usually ?
0
 
LVL 28

Accepted Solution

by:
Ryan McCauley earned 500 total points
ID: 40330746
Odd - I get an error when I try to view anything at oracle-base.com now, including the article I sent you in the previous post. I hope that's temporary, but I can't confirm what steps were included since I can't view the article at the moment.

Active dataguard isn't a failover solution - it's replication, where the second copy of the data is read-only. If you want to point read/write clients there in case your primary server goes offline, you've got to open the standby database for writes, meaning you'll break dataguard. To move things back to the primary server, you'd need to get it resyncronized once you're back online, which usually means just setting up dataguard the other direction and then cutting back the same way.

When you do a traditional cluster, I'm referring to multiple nodes that utilize shared storage but where only a single node owns the storage and has Oracle online at any given time. In those cases, failover takes the following amount of time:

1. Secondary node realizes first one is down (or, if you're failing over manually, the amount of time to dismount the service on the primary node
2. Storage arbitration time with the SAN to assign ownership to second node
3. Notifying the network switches of the new route for the cluster resource IP
4. Bring service online on the second node, including recovery of databases if necessary
5. Open service for client connections

In my experience, that's about 30 seconds of total time, but it depends on your hardware, network latency, and a few other things - it could be more or less time. When you're using RAC, both nodes are open and listening for connections at all times - the failover of a single node could result in a client connection being dropped, but the Oracle service will always be available and processing transactions.
0
 
LVL 1

Author Comment

by:marrowyung
ID: 40332289
"When you're using RAC, both nodes are open and listening for connections at all times - the failover of a single node could result in a client connection being dropped, but the Oracle service will always be available and processing transactions"

yes, and hte transaction rollback and wait for the other nodes, right ? how long it takes in your case?

"In my experience, that's about 30 seconds of total time, but it depends on your hardware, network latency, and a few other things"

that one is your active data guard failover time but not RAC, right?

any answer on that:

"
  As for the distance of RAC, I've heard as well that RAC isn't feasible over more than about 100km because of the latency you experience "

 any best practice on this  ?
  tks.

 also the link you show me is setting up VM box under linux and install Windows 2008 on that VM box but not the oracle is setup on Windows already.  or it just show me how to make RAC works on that on that windows VM which have oracle RAC installed on top of Windows ?

Open in new window

0
 
LVL 1

Author Comment

by:marrowyung
ID: 40338430
0

Featured Post

How to run any project with ease

Manage projects of all sizes how you want. Great for personal to-do lists, project milestones, team priorities and launch plans.
- Combine task lists, docs, spreadsheets, and chat in one
- View and edit from mobile/offline
- Cut down on emails

Join & Write a Comment

Cursors in Oracle: A cursor is used to process individual rows returned by database system for a query. In oracle every SQL statement executed by the oracle server has a private area. This area contains information about the SQL statement and the…
I remember the day when someone asked me to create a user for an application developement. The user should be able to create views and materialized views and, so, I used the following syntax: (CODE) This way, I guessed, I would ensure that use…
Via a live example, show how to take different types of Oracle backups using RMAN.
This video shows how to Export data from an Oracle database using the Datapump Export Utility.  The corresponding Datapump Import utility is also discussed and demonstrated.

744 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

Need Help in Real-Time?

Connect with top rated Experts

16 Experts available now in Live!

Get 1:1 Help Now