Newbie cluster question: where does the admin server live?

Hello, I'm looking at clustering for our application where I work, and so I'm reading through the cluster-related documentation for 11g. I have a sort of an architecture question: where should the admin server be started? Hypothetically speaking, if you have two nodes for your cluster, would the AdminServer run on a 3rd box which sort of stood in front of the two nodes?
Thanks much -

The ideal situation would be for your admin server to be separate from the machines hosting your managed server. This allows you to control access to the admin server, eliminate the performance impact if you have the admin + managed on the same host, and limits impact if youir #1 host fails, etc.
But companies may be unwilling to invest in a distinct (3rd) host just for the admin server, especially if you have multiple environments ( prod, testing, dev, etc. ).
So usually the admin winds up sharing the host with a managed server.

Similar Messages

Where does the ACS server get the DNS info for IP pools?

I am trying to change the DNS servers that my VPN users are assigned from the IP pools on the ACS server. Where does the IP pools get the DNS server information. I have changed the DNS Ips on the windows server and rebooted. But the VPN clients are still being assigned the old DNS servers.

ACS ip pools do not push DNS server information
It is either being forwarded from VPN concentrator group setup or
it is being send from ACS user/group setup > Radius(VPN 3000) attributes > [026/3076/005] Primary-DNS.
Hope this helps.
Regards
Rohit

Strange WLS6.sp2 Oracle DB Reconnect issue in a Cluster but ONLY on the Admin server

Any Ideas? ->
I am having a issue when our oracle db server goes down and when we try
to reconnect the Pool via weblogic I am getting this error. This only
happens on the Admin server for the cluster, none of the managed servers get
this and they all seem to be ok. Any ideas?
-Pete
**** Below is the stack trace:
java.sql.SQLException: Io exception: Broken pipe
at oracle.jdbc.dbaccess.DBError.throwSqlException(DBError.java:114)
at oracle.jdbc.dbaccess.DBError.throwSqlException(DBError.java:156)
at oracle.jdbc.dbaccess.DBError.throwSqlException(DBError.java:269)
at
oracle.jdbc.driver.OracleStatement.executeNonQuery(OracleStatement.java:1329
at
oracle.jdbc.driver.OracleStatement.doExecuteOther(OracleStatement.java:1232)
at
oracle.jdbc.driver.OracleStatement.doExecuteWithBatch(OracleStatement.java:1
353)
at
oracle.jdbc.driver.OracleStatement.doExecute(OracleStatement.java:1760)
at
oracle.jdbc.driver.OracleStatement.doExecuteWithTimeout(OracleStatement.java
:1805)
at
oracle.jdbc.driver.OraclePreparedStatement.executeUpdate(OraclePreparedState
ment.java:322)
at
oracle.jdbc.driver.OraclePreparedStatement.execute(OraclePreparedStatement.j
ava:366)
at
weblogic.jdbc.pool.PreparedStatement.execute(PreparedStatement.java:45)
at
weblogic.jdbc.rmi.internal.PreparedStatementImpl.execute(PreparedStatementIm
pl.java:289)
at
weblogic.jdbc.rmi.SerialPreparedStatement.execute(SerialPreparedStatement.ja
va:398)

Hi,
As i understand is OS is clustered not the db. Better is to manually shutdown the db clean and then start it.
TO start login through sqlplus sys as sysdba
1. startup
2. Once its done, start the listener
3. Start the application
Anand

(SOLVED) where does the xorg server gets the Composite file from?

I'd like to see the list of composite combinations. Seeking on the Internet I understand that the file should be:
/usr/share/X11/locale/en_US.UTF-8/Compose
Where en_US.UTF-8 is my locale, but according to that file:
<Multi_key> <underscore> <o> : "ō" U014D # LATIN SMALL LETTER O WITH MACRON
I should obtain the small letter o with the macro using multi_key underscore and o. Instead I obtain this: º.
Same problem with the
<Multi_key> <c> <o> : "ǒ" U01D2 # LATIN SMALL LETTER O WITH CARON
If I press the keys I obtain: ©, the copyright sign...
So... Where does the arch xorg got its Compose file? Change I change its behavior?
Thanks.
Last edited by ezzetabi (2009-10-10 15:25:02)

I ignored that is was an application side issue, thanks for explanation. And indeed I was using lxterminal: the terminal emulator of lxde that is based on GTK. Using the xim input system everywhere solved the problem.
People! Listen! Put this two lines in your .xinitrc :
. /etc/X11/xinit/xinitrc
export GTK_IM_MODULE=xim
export QT_IM_MODULE=xim
exec whatever you use
Thanks! Cheers!
Last edited by ezzetabi (2009-10-10 15:25:25)

P6N Diamond, SLI question: where does the 2nd card go?

As per title, I have a P6N Diamond motherboard with 4 video card slots (four PCI-e x16), two yellow + two white. I have a single GPU/card in the first slot as usual, but I got a 2nd GPU/card today. Where does it go so I could get 16x16 SLI? Manual is unclear.
2nd white slot (third PCI-e x16) or the yellow (second PCI-e x16) ?
I ask because this motherboard has non-standard color coding. Even the RAM slots are not color coded (channel-wise). Instead of A-B-A-B color arrangement on the RAM slots as every other motherboard out there, this motherboard has A-A-B-B arrangement.

Update: First of all, the case connectors on the motherboard are in the way for SLI purposes. They are too tall and block the insertion of the second card. On my old ASUS moterboard the case pins were on the bottom of the PCB, out of the way, but on this MSI motherboard they are on the side, lined up with the PCI-e x16 slot. Really dumb design! It might be fine for short video cards like 9600GT or similar, but for extended length dual slot video cards like GTX260 (which what I'm using) they are in the way, so much that the second card can't be inserted into the slot at all.
I spent two hours modding the connectors. I had to snap and break the plastic plugs, remove the metal pins and mod them so they are only 3mm tall, instead of 9mm. I know this motherboard is old so MSI couldn't plan for GTX260/280 and other long cards, but cards like 7950GX2 and similar did exist and predate this motherboard so there is no excuse for MSI for placing the connectors in that position.
Anyways, warning for any P6N Diamond owners: if you have extended length or longer dual slot video cards like GTX 260, GTX 280, and 4870x2 you will have problems with clearance and would need to do some severe modding or get a different motherboard.

Newbie M5 question - How does the Censor effect blur only the inside of the circle?

I'm brand new to Motion 5. I'm trying to understand how an effect like "Censor" applies the blur to only the inside of the circle. I opened the Censor effect in M5 and studied it, but the answer is just out of my grasp. Can someone explain it to me, or point me to an article that will help? Thanks.
Dan

In Motion, in the Censor effect project, if you turn off the Effect Source (lowest layer) you'll see the "circle" used to "limit" the effect.
Actually, this is just an illusion:
The original Effect Source (in FCPX, that's whatever clip this effect is applied to) is Cloned (the layer is called "Mask"): a clone is basically a "rasterized copy" of the original -- what I mean by that is that Clones not only copy original objects, but also all their filters and other layers/effects if you clone a group of objects -- reducing them all down to a single layer copy. This copy can be treated just like any other object in Motion (with an interesting exception*): you can add masks and other filter effects and/or behaviors to create new actions while the original source of the clone is left intact.
In this effect, there is a Circle Mask applied to the Clone, as well as three different Filters: Pixellate, Gaussian Blur, and Brightness. These three different filters are turned on and off separately by use of a "pop-up" rig so that only one of the filters is applied to the clone at a time. The gear shaped objects are Behaviors, in this case Link behaviors that "link" to the Circle Blur filter properties that is attached to the Effect Source which control the location in the canvas and the scale of the mask (Radius to Scale-X and another to Scale-Y). The Circle Blur filter has Onscreen Controls of a center point and a "radius" and those property values can be shared to other objects in a project via these link behaviors. The Circle Blur filter effect is turned off in this project because it *only* needed as a *source* for the onscreen control properties (and those can be turned on for use in FCPX [Publish OSC] whether the filter type is used or not.) All of the visible action in this project is created by applying other filter effects to the Clone copy and masking the area so that the rest of the source is uneffected while this effect (Censor) is active in FCPX.
The *interesting exception: Clones can be "retimed" -- even timed to go in reverse. No other object (other than actual video) can be made to do the same. What does this mean? It means that you can create, basically, an "in-project" video of other objects in animation (without exporting the animation first) and retime them to loop over and over, go in reverse, speed up, slow down, etc... Clones are a very useful tool in Motion (not to mention "light-weight"). Their only drawback: they *flatten* 3D animations, so you can't apply "more 3D" effects to them because they'll just look flattened on a plane if you do. So any time you feel you need to duplicate something, ask yourself if you would be better served by a clone instead.
HTH

When I share a file to YouTube, where does the output file live? I want also to make a DVD. And is this a best practice or is there a better way?

I want also to make a DVD, but can't see where the .mov files are.
And is this a best practice or is there a better way to do this, such as with a master file?
thanks,
/john

I would export to a file saved on your drive as h.264, same frame size. Then import that into youtube.
I have never used FCP X to make a DVD but I assume that it will build the needed vob mpeg 2 source material for the disk.
I used to use Toast & IDVD. Toast is great.
To "see" the files created by FCP 10.1.1 for YouTube, rt. (control) click on the Library Icon in your Movies/show package contents/"project"/share.

Where does the system preferences live? mine disappeared by the finder bar

i've lost my system preferences from my finder bar. please tell me how i can retrieve them.
thanks,
brad

Do you mean it's gone from the Dock (bottom of the screen) or from the apple menu in the top right-hand corner?
If you mean it's no longer in the Dock, just open your Applications folder and drag the System Preferences icon into the Dock.

6.0 cluster startup without the admin server...

Is it possible to start a WL6.0 cluster without the admin server being up
          and running? If not, are there any workarounds that eliminate the admin
          server as a single point of failure?


          I have same questions.
          1.The admin server's primary task is up the cluster?And if admin server is down
          that can't affect the cluster running.
          2.Is it right when the cluster is working,the useful ip is admin server's ip?For
          example,admin 10.10.10.1,cluster1 10.10.10.2,cluster2 10.10.10.3,the only useful
          ip is 10.10.10.1 in the cluster.
          3.Is it correct that because the only useful ip is 10.10.10.1,therefor my client
          access the cluster from 10.10.10.1?
          When my admin server is down,then my system will be down.
          4.which is the dispatch in the cluster system ?
          my step of configuration of th cluster:
          1.add two managed servers to admin server and edit startweblogic.cmd;
          2.config cluster.(admin server address 10.10.10.1:8168,cluster address 10.10.10.2:7001,10.10.10.3:7001),multicast(237.0.0.1)
          please help explain these questions!thanks a lot!
          2.
          Kumar Allamraju <[email protected]> wrote:
          >Sorry if i wasn't clear, what i mean to say is
          >
          >When you first start the managed server's you have to have admin server
          >up &
          >running.
          >Admin server can go down when cluster is running, but then you can't
          >add servers
          >to cluster
          >and I think its ok to have it that way from administration point of view.
          >
          >Yes, there's no facility to achieve single point of failure for admin
          >servers.
          >
          >--
          >Kumar
          >
          >
          >
          >William Wallace wrote:
          >
          >> I may be misreading that page, but it doesn't really tell me how to
          >> successfully start to start a managed server is the admin server is
          >down. If
          >> you run the managed server script you get an "cannot connect to admin
          >> server" exception if the admin server is not running. Now, I imagine
          >that I
          >> could script something up to restart the admin server if it's down,
          >but that
          >> doesn't protect me from hardware failure.
          >>
          >> "Kumar Allamraju" <[email protected]> wrote in message
          >> news:[email protected]...
          >> > http://e-docs.bea.com/wls/docs60/adminguide/startstop.html#1035064
          >> >
          >> > --
          >> > Kumar
          >> >
          >> >
          >> > William Wallace wrote:
          >> >
          >> > > Is it possible to start a WL6.0 cluster without the admin server
          >being
          >> up
          >> > > and running? If not, are there any workarounds that eliminate the
          >admin
          >> > > server as a single point of failure?
          >> >
          >

Cann't connect to the Admin Server!

Hi,everyone:
when I "startconsole",error appears:
"Cann't connect to Admin Server:http://myhostname.mydomain:port number,the URL is not correct or the server is not running "
how can I resolve this problem?
Thanks

This may seem like a dumb question, but
Is the Admin Server running on myhostname.mydomain:port number?

Where on the HD does the iPhoto library live?

Hi, I'd like to copy photos to an external HD. Where does the iPhoto library live on my hard drive? It's not in the "Home" (picture of a house) directory (there is a "Pictures" folder there, but my iPhoto pics are not in here). Where else could it be?

James, the search took you to iPhoto.app?
Here is a screenshot of the iPhoto Folder and where it resides as default.
This is a view of how the iPhoto Library looks in the Finder.
It is organized by Year folders. Inside the Year folders are Month folders. Inside the Month folders are the days of the Month. Inside the day folders are the photos you took on that day. If you imported video clips and RAW files on that day they will be in an Originals folder. If you have edited any photo imported on that day, the original will be in that folder also.

Does SNMP agent on the admin server polls managed objects automatically?

Hi All,
If I enable SNMP agent on the admin server, does it mean that it polls and caches MIB of all the managed servers in the domain and return it if a client requests for information or it polls and caches that MIB value only when requested?
2)Suppose a client requests for SNMP data of a managed instance and the instance is down, does admin server keep polling or does it timeout after a number of secs or retries. I don't see any config params in weblogic 8.1.3 or weblogic 9.0 console.
Thanks. I appreciate your answers.
Thanks,

Are you saying the notes are incorrect?
No, I wasn't commenting on that at all. I was just saying that DBUA is the recommended way of upgrading databases... and that's in spades if you're talking about 100GB databases.
I'll lay odds that DBUA to 10.2.0.1 and then patching to 10.2.0.3 is quicker than exporting and importing 100GB of data.
(Applying a 10.2.0.3 patch is done with opatch and doesn't require a fresh installation of Oracle. A patch mostly just modifies the data dictionary: it's finished in minutes, not hours).
My real point: these sorts of questions are incredibly easy to ask, answer, discard, re-ask, modify, mull over and basically experiment with in a virtual environment.
And again: it's utterly bizarre why you wouldn't have physical access to this database (and keep in mind that I can't keep in mind little nuggets of information such as this one which you didn't see fit to share earlier!)
It's in a virtual server. Virtual servers can be zipped up and emailed! (OK, maybe not 100GB... but I've shipped a small hard disk in a zip-lok pack envelope half way round the world before now -though obviously it wasn't critical to have the contents of that envelope available for use all the time... Still, it sounds to me like here's a technology (virtualisation) that's not being properly exploited).

Where does the BPEL project goes when deployed to a Linux server?

Hi,
Where does the BPEL project goes (I need the path) when deployed from JDeveloper into a Linux server?
Also I would like to know the file name. We have been searching for a *.jar file but we cannot find it.
Please help.
Amorsolo

And also, one zip file with size 0KB get created under$ORACLE_HOME$ \integration\esb\oraesb\deploy.
This will be named *'bpelservice_+<yourdomainname>+_<processname>'*
Edited by: Arjun_M on Sep 10, 2009 6:13 AM
Edited by: Arjun_M on Sep 10, 2009 6:13 AM
Edited by: Arjun_M on Sep 10, 2009 6:13 AM

Does NodeManager make the Admin Server restart at reboot?

I have been trying to run both my Admin server and SOA Suite using the Node Manager based on this article:
http://blogs.oracle.com/jamesbayer/2010/01/weblogic_nodemanager_quick_sta.html
When I reboot, the soa_server1 (managed server) starts up but the AdminServer does not.
Am I missing something? Is this not something the NodeManager is supposed to support?
Thanks for quick replies,
Jeremy

It depends on how you have deployed your application on Weblogic. Generally, organizations do not have any of their applications on the Admin server. They generally use a clustered set of managed servers to have their applications deployed on them. From the security perspective also, the Admin server is preferred to be kept down as this prevents unauthorized users from gaining access to the Admin console and making changes on them. However, before keeping the Admin server down, you must make sure that none of your application components are deployed on the Admin server or make use of the Admin server.
Edited by: Smithan John on Jun 8, 2010 1:33 PM

Weblogic does not support failover for the admin server.

There is a single point of failure in the current weblogic architecture. If the
admin server dies there is no automatic alternate admin server.

That is true, and we plan to fix it in one of the next major releases. Currently
the story is not as bad as it appears to be for following reasons:
1. Failure of admin server doesn't mean failure of already running managed server.
2. You can start a new server as admin server of domain with discover flag=true and
it will recognize all already exisiting managed server.
Viresh Garg
Principal Developer Relations Engineer
BEA Systems
naresh gupta wrote:
There is a single point of failure in the current weblogic architecture. If the
admin server dies there is no automatic alternate admin server.

Newbie cluster question: where does the admin server live?

Similar Messages

Maybe you are looking for