Channel Bonding? Oracle RAC with multiple redundant network

We are planning on setting up on Oracle RAC 10g on Linux.
We have redundant switches, so I would like to set up the network to have some version of network redundancy.
Under windows, it is called 'teaming'. Linux is 'channel-bonding'
I just want to make sure if I setup the channel-bonded interfaces, it will be supported with RAC. Has anyone done this before??
-Andy

we are running our RAC on NIC bonding. Although we need to do some more testing to be real sure as to the functionality of the bond.
Metalink Note:298891.1 would be a good place to start.
hth,
-S

Similar Messages

SQLLoader issues of Oracle RAC with ASM

One of our client wants to use Oracle RAC with ASM for our application, i just want to know if there would be any two-phased commit transactions and SQL*Loader issues in ASM.
Database is Oracle 10g

ASM works only at storage layer and has nothing to do with:
- distributed transactions
- client executable that connects to database instance: SQL*Loader, SQL*Plus, etc.
RAC has also nothing to do with distributed transactions: a RAC database is a single database with multiple instances but still a single database; there is no need to use distributed transactions because you have a RAC database.
Edited by: P. Forstmann on 24 févr. 2011 13:27
Edited by: P. Forstmann on 24 févr. 2011 13:31

How to investigate high io in a RAC with multiple databases?

how to investigate high io in a RAC with multiple databases?

RAC is saying that a single server lacks the processing capacity needed for databaseA. So additional databaseA instances on additional servers are needed in order to meet the processing demands for databaseA. Thus RAC is required.
Multiple databases on a single server says that the server has spare processing capacity and that it can not only deal with databaseA processing demands, but also with the processing load of databaseB.
So how does your statement about running multiple database RAC instances on the same RAC server node, make sense? And does your I/O issue not immediately point to the fact that no, RAC server1 cannot deal with the I/O resources required by both databaseA and databaseB?
After all, how is I/O scaled on RAC? How is ANYTHING scaled on RAC? By adding more databases to a RAC server node? Of course NOT!!
RAC scales by adding more servers (with a new database instance each) to a physical database in order to scale processing capacity and I/O. By adding more I/O channels to access the SAN/shared storage. By adding more CPUs to increase processing capacity.
Bluntly put - multiple RAC database instances on a single RAC server node is kind of an idiotic approach to scalability in my view....

Does NFS support Oracle RAC with TAF ?

Hi,
Does NFS support Oracle RAC with TAF feature. ?
If yes, please point to some valid document to prove the same.
Regards
Sumit

TAF has nothing to do with the underlying storage architecture and is therefore 'supported' with your choice of asm, raw devices, cluster file system and NFS. NFS is supported on some platforms. For more details, check the certify tab in metalink. Some information can also be found here
TAF will be available on any rac installation, but not all drivers support it (jdbc-thin for example)
Bjoern

Oracle RAC with more then two nodes?

Hello,
knows anybody a reference client project that use Oracle RAC with more then two nodes on a Linux environment?
Many thnaks!
Norman Bock

Hello Norman,
XioTech is a SAN company that has a project called "THE TENS". They configured and ran a 10 node RH Linux Oralce 9i RAC. I understand they want to see if 32 are possible. I am sure if you ask them, they would be happy to give you the details.
http://www.xiotech.com/
Cheers

RAC with Multiple controlfile's

Is it possible to start multiple oracle instances with different controlfiles (with one database name/alias) instead of using the common controlfile.
right now we are using all files in common except pfile like standalone pfile for each instance.
The way we are able to control the database with pfile/like block size and some other common parameters.
in the same way does the oracle 10g or any future database support the separate controlfile for each instance.
DBA should be able to choose the controlfile parameters/datafiles .

user538769 wrote:
I have 7 application systems that need 7 different Oracle database servers. It is not that simple. There is no such "+rule of thumb+" even that states that one should have one application system per database. So you're assumption that because you have 7 app systems you need 7 Oracle db servers, is very much flawed.
There are numerous factors that dictate how resource intensive that app system is and the size of the server it needs (or the number of servers in terms of RAC).
How many nodes are to be configured to support 7 instances of database?7 what instances? 7 dedicated non-RAC instances? It is seldom technically sound to run 7 (or more) database instances on a single server platform. Often better performance and scalability are achieved by consolidating those difference instances into a single instance.
7 RAC instances? A RAC instance per server is needed. Thus 7 servers would be required.
But this is IMO still very much irrelevant thumb suck equating 7 apps to need 7 RAC instances.
This is not how client-server architecture works. It is not a hardware architecture - it is a software architecture. For that reason it is wrong to assume that each s/w layer or component in the client-server architecture needs a separate and dedicated h/w platform.
Client-server does not scale using hardware alone. In fact, scalability via h/w alone can be a dead-end street running into a massive and unmovable wall.

Oracle RAC with ASM install failure

Hi Guy's
I've just rebuilt an Oracle RAC system, i've built this system many times and have the build documented. I usually follow the build doc to the tee and everything works fine. I have now been asked to rebuild the system to go into the production evironment. This is a two server standard RAC setup using ASM. We have a HP MSA500 shared storage setup. I've run the cluster verification tool and all the checks have come out fine. Whilst installing the clusterware i get a failed error message whilst it try's to confiure the Oracle Clusterware Configuration Assistant.
The error message i get is:
Command = C:\Windows\system 32\cmd\C Call E:\Oracle\Product\10.2.0\crs\install\crssetup.config.bat
PROT-1:Failed to intialize OCR Config
STEP 1 Checking status of CRS cluster
STEP 2 Creating directories (E:\Oracle\Product\10.2.0\Crs
STEP 3 Configuring OCR repository
ocr upgrade failed with (-1)
I've done this build many times and never seen this error. From what i've read on the internet it looks like it could be something up with my shared storage but i dont know what.
Any help would be greatly appreciated
Thanks
Lee

try to apply ''dd' command against OCR & Voting disks to ensure that their headers are formatted 100% and contains no data from old installation:
dd if=/dev/zero of=YOUR_OCR_DISK bs=1024 count=10000
dd if=/dev/zero of=YOUR_VOTE_DISK bs=1024 count=10000

Oracle RAC with Admin tools

Hello!
I use Weblogic Server 10.3.2. I have a portal application and I want to manage the content with Portal Administration Console from admin-tools.war . Admin-tools is a part of my portal application. I eclipse I check for my ear project on Project Facets : Admin Console, Admin Framework, Portal Application Services, Portal Customizations Framework.
I have a cluster with two servers. As a database I use Oracle 10 with two instances : instance 1, instance 2. I have created my domain with wizard and I checked RAC Multidatasources.
All datasources and multidatasources are targeted to my weblogic cluster, also where is deployed my app and Portal Administration Console ( admin-tools.war).
I use the multidatasources :
Multidatasource Name : cgDataSource JNDI:      cgDataSource     Load-Balancing
Oracle Driver for data sources cgDataSource-rac0, cgDataSource-rac1 : oracle.jdbc.xa.client.OracleXADataSource
MultiDataSource Name : cgDataSource-nonXA     JNDI : cgDataSource-nonXA      Load-Balancing     Oracle Driver:
Oracle driver for datasources from multidatasource : oracle.jdbc.OracleDriver . Supports Global Transaction
Name : p13nDataSource
JNDI : p13n.trackingDataSource,p13n.sequencerDataSource,cm.sequencerDataSource,p13n.leasemanager,p13n.dataSyncDataSource,p13n.entitlementsDataSource,p13n.jobManagerDataSource,p13n.quiescenceDataSource,p13n.credentialsDataSource     Load-Balancing
Oracle driver for datasources from multidatasource: oracle.jdbc.OracleDriver .
!!DOESN'T Supports Global Transaction!!
Name: portalDataSource
JNDI : weblogic.jdbc.jts.commercePool contentDataSource contentVersioningDataSource portalFrameworkPool     Load-Balancing
Oracle driver for datasources from multidatasource: oracle.jdbc.OracleDriver .
Supports Global Transaction
Name: portalDataSourceAlwaysXA
JNDI : portalFrameworkPoolAlwaysXA     Load-Balancing
Oracle driver for datasources from multidatasource: oracle.jdbc.xa.client.OracleXADataSource
Name: portalDataSourceNeverXA JNDI: portalFrameworkPoolNeverXA
Oracle driver for datasources from multidatasource: oracle.jdbc.xa.client.OracleXADataSource
But when I try to add a content, after a server restart, I receive the following error :
<Error> <PortalAdminTools> <sun1> <Server-0> <[ACTIVE] ExecuteThread: '5' for queue: 'weblogic.kernel.Default (self-tuning)'> <weblogic> <> <> <1289472419642> <BEA-415097> <A RepositoryException was thrown: java.sql.BatchUpdateException: ORA-00001: unique constraint (WEBLOGIC_PORTAL_PROD.PK_CM_PROPERTY) violated
com.bea.content.RepositoryException: java.sql.BatchUpdateException: ORA-00001: unique constraint (WEBLOGIC_PORTAL.PK_CM_PROPERTY) violated
     at com.bea.content.internal.server.common.dbaction.Batcher.processWithRetries(Batcher.java:148)
     at com.bea.content.internal.server.common.dbaction.Batcher.processAnyRemainingItems(Batcher.java:86)
     at com.bea.content.repo.internal.server.dao.PropertyDao.insert(PropertyDao.java:383)
     at com.bea.content.repo.internal.server.persister.JDBCNodePersister.createNode(JDBCNodePersister.java:148)
     at com.bea.content.repo.internal.server.logic.NodeOpsLogic.create(NodeOpsLogic.java:227)
     at com.bea.content.repo.internal.server.bean.NodeOpsBean.create(NodeOpsBean.java:106)
     at com.bea.content.repo.internal.server.bean.RepoNodeOps_ehgg7y_ELOImpl.create(RepoNodeOps_ehgg7y_ELOImpl.java:323)
     at com.bea.content.repo.internal.client.common.Node.create(Node.java:123)
     at com.bea.content.spi.internal.NodeOpsImpl.createNode(NodeOpsImpl.java:823)
     at com.bea.content.spi.internal.ExtendedNodeOpsImpl.createNodeWithNodeTypeAndObjectClassAndWorkflow(ExtendedNodeOpsImpl.java:81)
     at com.bea.content.federated.internal.delegate.LatestInterfaceVersionWrapper.invoke(LatestInterfaceVersionWrapper.java:72)
     at $Proxy457.createNodeWithNodeTypeAndObjectClassAndWorkflow(Unknown Source)
     at com.bea.content.federated.internal.filter.logging.NOPSLoggingFilter.createNodeWithNodeTypeAndObjectClassAndWorkflow(NOPSLoggingFilter.java:111)
     at com.bea.content.federated.internal.delegate.NodeLogic.createNodeWithType(NodeLogic.java:275)
Caused By: java.sql.BatchUpdateException: ORA-00001: unique constraint (WEBLOGIC_PORTAL.PK_CM_PROPERTY) violated
     at oracle.jdbc.driver.DatabaseError.throwBatchUpdateException(DatabaseError.java:629)
     at oracle.jdbc.driver.OraclePreparedStatement.executeBatch(OraclePreparedStatement.java:9447)
     at oracle.jdbc.driver.OracleStatementWrapper.executeBatch(OracleStatementWrapper.java:211)
     at weblogic.jdbc.wrapper.PreparedStatement.executeBatch(PreparedStatement.java:191)
What is the problem? Please tell me if you need more details about my configuration and problem.

This issue is beyond simple WLS JDBC. I suggest opening a WLS portal bug.

Oracle RAC with Oracle Linux Based RAC Applications

What are the advantages of using Oracle Linux Based RAC applications (also known as Linux Oracle Projects) with Oracle Database RAC 10g ?

Personally, I find RAC to be about 1000% easier to manage on Linux...than on what? And why?
Beacause personally I can't see difference in RAC management between Linux and Solaris (i.e.).

RAC with multiple databases

hi,
we are evaluating a design for the configuration of 3 databases in 10g RAC in linux
Available HW: 2 servers each with 32 gb ram
No of databases need to be configured :3
what would be the effective configuration?
the options we think are,
1) One 2 Node RAC with 3 databases or
2) Three 2 Node RAC each with 1 database each
what would be the licencing cost in different options
please help
tia,
newbie
Edited by: user570138 on May 26, 2009 11:29 PM

user570138 wrote:
hi,
sorry, am new to RAC.
heard about one implementation wherein with a h/w of 2 servers,
1) 2 node cluster for one database
2) single node cluster for the second database in one of the servers.
is it possible to configure multiple clusters on the same machines?
here in our case, we have 3 databases that need to part of RAC configuration.A single node cluster? A "cluster" implies two or more servers.
If you have two servers, you can create two node cluster for any number of RAC database(s).
Also you can create stand-alone database(s) on either server.
If you use VM, you can cluster as many VM machines as you have configured.

Oracle RAC with QFS shared storage going down when one disk fails

Hello,
I have an oracle RAC on my testing environment. The configuration follows
nodes: V210
Shared Storage: A5200
#clrg status
Group Name Node Name Suspended Status
rac-framework-rg host1 No Online
host2 No Online
scal-racdg-rg host1 No Online
host2 No Online
scal-racfs-rg host1 No Online
host2 No Online
qfs-meta-rg host1 No Online
host2 No Offline
rac_server_proxy-rg host1 No Online
host2 No Online
#metastat -s racdg
racdg/d200: Concat/Stripe
Size: 143237376 blocks (68 GB)
Stripe 0:
Device Start Block Dbase Reloc
d3s0 0 No No
racdg/d100: Concat/Stripe
Size: 143237376 blocks (68 GB)
Stripe 0:
Device Start Block Dbase Reloc
d2s0 0 No No
#more /etc/opt/SUNWsamfs/mcf
racfs 10 ma racfs - shared
/dev/md/racdg/dsk/d100 11 mm racfs -
/dev/md/racdg/dsk/d200 12 mr racfs -
When the disk /dev/did/dsk/d2 failed (I have failed it by removing from the array), the oracle RAC went offline on both nodes, and then both nodes paniced and rebooted. Now the #clrg status shows below output.
Group Name Node Name Suspended Status
rac-framework-rg host1 No Pending online blocked
host2 No Pending online blocked
scal-racdg-rg host1 No Online
host2 No Online
scal-racfs-rg host1 No Online
host2 No Pending online blocked
qfs-meta-rg host1 No Offline
host2 No Offline
rac_server_proxy-rg host1 No Pending online blocked
host2 No Pending online blocked
crs is not started in any of the nodes. I would like to know if anybody faced this kind of a problem when using QFS on diskgroup. When one disk is failed, the oracle is not supposed to go offline as the other disk is working, and also my qfs configuration is to mirror these two disks !!!!!!!!!!!!!!
Many thanks in advance
Ushas Symon

I'm not sure why you say QFS is mirroring these disks!?!? Shared QFS has no inherent mirroring capability. It relies on the underlying volume manager (VM) or array to do that for it. If you need to mirror you storage, you do it at the VM level by creating a mirrored metadevice.
Tim
---

Oracle RAC with operating system virtual network cards

Hi,
I have two servers with four network cards each, two for public and two for private network. Each public card has own IP-address and the both have the virtual IP-address (from OS, not from Oracle Clusterware). Each private card has own IP-address and the both have the virtual IP-address. I have full redundancy for network cards. How should define Oracle Clusterware VIP-address for each machine?
Can somebody write me sample /etc/hosts file for this configuration?
Thanx,
Jacek

Oracle RAC and crossover cable for private network

Problem with Oracle RAC with DRCP and persistent connections

Hello,
I am having a problem that has me stumped. I cannot get SCAN, and DRCP to work together nicely.
Right now, I can connect through my scan hostname fine without connection pooling, once I enable it I start having connection time outs. I found if I specified outbound_connection_timeout in sqlnet.ora that it times the connection out to the time I set in there. If I don't set this option I get a "ORA-12170: TNS:Connect timeout occurred".
The odd thing is I can leave connection pooling on and connect to one of my rac instances individually and it doesn't experience this problem. The only time it pops up is if I combine multiple rac instances (Through the use of SCAN or tnsnames.ora) and I enable connection pooling.
I don't know which direction to go to even diagnose this, so any and all help is appreciated.
Thank you.
Edited by: Rarp on Oct 1, 2011 10:47 PM

Hi,
You need to detail the problem. Enable TRACE SQL*NET on the server and client using note below.
How to Enable Oracle SQLNet Client , Server , Listener , Kerberos and External procedure Tracing from Net Manager [ID 395525.1]
Also you can try check this note:
11g: ORA-12170 With Combination RAC, DRCP and a Firewall [ID 953277.1]
Hope this helps,
Levi Pereira

How to write to one channel of a task with multiple channels? (plus other things...)

So I have a USB-6009 DAQ. It has 12 digital output lines. I want one channel that is "Dev0\line0:10" which represents an address bus in my application. And a second channel "Dev0\line11" which represents a program enable line in my application.
I have tried creating two different tasks and adding one channel to one task and the other channel to the other task. The only task that worked was the task with "Dev0\line0:10". It was always the task containing that channel, regardless of the order of creation. So then I moved on to a different method. (I read somewhere that I should only create one task of each type. like only 1 DO task only 1 AO task etc... However I am also using the two analog ouputs to and have a task for each AO and they work just fine.)
I tried add both channels to one task. But when I needed to control only the address bus, I had to have some information for the other channel as well. This was a little trying, but I could configure it that way. But it turned out to be easier for me to just make one channel with all the lines and OR in the data for line11 with each write.
I just wondered if it was possible to write to one particular channel of a task and not the other channel? That would really be the ideal solution for me. especially if I could write multiple samples to the one channel while I left the other channel alone. Which brings up another complaint... Why does WriteMultiSamplePort only work if I use a DigitalMultiChannelWriter, and not if I use a DigitalSingleChannel writer even though my task only has one channel (and by the way I set up the channel as one channel for all lines)? A perplexing issue to be sure.
And no I can't just load all my samples into an array and write them all at once because I also have to manipulate the two analog outputs in between the various digital writes.
I am using NI DAQmx 7.5 and C#. I am trying to use the DAQ to program a digital switch, which has proven to be a real challenge. In push-pull mode there is too much ringing for the switches programming port to tolerate. But the switches interface is LVTTL, so I needed the 3.3V. When I changed to open-collector I had to use voltage dividers to drop down to 3.3V. But the rise time using open-collector is too slow for me the program the switch in serial mode, so I had to change to parallel. The switch has an 11 bit multi-plexed ADDR/DATA bus. So the DAQ I had chosen to use which had plenty of lines for the serial programming, now is strained to it's absolute limits by the parrallel interface. ARGH. The only output I am not currently using is the counter, and I'm going to need it if I ever want to read back from the switch. But first I have to seperate the ADDR/DATA bus from the CS line on the DIO lines of the DAQ. And I don't know for sure what I'm going to do about the voltage level translation when I have to go bi-directional. Maybe I can filter out the ringing in push-pull mode? Any thoughts on that?

Hi Saikey,
In most cases, you are exactly right: you can only use one task for one type of operation (i.e. only one analog input task in the same program). With the USB 6009, you can have multiple digital output tasks running at the same time. I was able to run a digital output program with two different digital output tasks configured for a USB-6009.
However, you stated that it would be better if you had everything in a single task for your application and write data to only some of the channels. The easiest way to do this is to modify your array of output data so that only the data to that one channel is changing. So, for example, if you keep writing a 0 to the channels that do not need new data nothing will change.
If you have to change your analog outputs during this program, you could create an event structure that would stop and restart the analog output tasks without changing the digital output data. I hope that you find this information helpful.
Regards,
Hal L.

Oracle, Issue with Multiple transaction from two nodes.

Hi All,
Below is the scenario ,that we are encountering.
We have Client which is RAC modelled having Node1 and node2 and the server having primary and secondary nodes.
From Client we are getting multiple transactions at a rate of nano-milliseconds to the server.
Our code setup:
1) Requests would be fired from client for either insert or update into DB to a Table(Table_A).
If record exists, we are going for update or else we are inserting.
Table Structure:
Table_A Table has Primary key on Account_id
Our Scenario is that
Scenario 1:
a) Database is verifying ,if the data exists in table.
b) Data not available(record count=0) in the table and as per functionality it goes for insert query.
c) Before commit occurs, another requests comes in and checks in table for data existence.
d) Since commit is yet to happen, the count on table is ZERO and hence the second request is also coming as INSERT.
e) Since two insert on same record comes in, Unique Key Constraint Violation error is been triggered.
Scenario 2: We tried to simplify the above process by using Merge option In Oracle. But even that is not solving our concern.
Thanks

Scenario 1 - you will probably have to perform some sort of locking to prevent this issue. However, doing so will hurt scalability.
Scenario 2 - your situation sounds perfect for the MERGE statement. Can you provide more details on "But even that is not solving our concern."? What are you doing and what are you seeing?
Also, post your version.
Cheers,
Brian

Channel Bonding? Oracle RAC with multiple redundant network

Similar Messages

Maybe you are looking for