IOM connections in a cluster setup

I just wanted to verify that I have my physical connections setup:
If I have 2 Fabric interconnects and 2 chassis, I should connect:
Chassis1-IOM1 connected to Fabric1
Chassis1-IOM2 connected to Fabric2
Chassis2-IOM1 connected to Fabric2
Chassis2-IOM2 connected to Fabric1
Is this correct? I just want to make sure I'm not supposed to connect it this way:
Chassis1-IOM1 connected to Fabric1
Chassis1-IOM2 connected to Fabric2
Chassis2-IOM1 connected to Fabric1
Chassis2-IOM2 connected to Fabric2
The Cisco high availability http://www.ciscosystemscapital.biz/en/US/products/ps10278/products_configuration_example09186a0080aef90f.shtml document doesn't specify how to connect 2 chassis. Thanks.
Thanks.

Hi Terence.
     The first thing you need is to define is how many connection you want from the Chassis to each of the Fabric Interconnects.
     IO Modules connect to either FIC A or B (not to both) - If you connect port 1 of the left IO Module to FIC A then the other ports also need to connect to it .
     Then you should have all your chassis connected all the same way (ideally though you may have more bandwidth requirements in one chassis then on the other and therefore have more connections - server ports on one chassis than on the other).
     The above means that if you are using only 1 server port per FIC to the chasssis you should connect each chassis like this (for example):
     Chassis 1 IO Module 1 (left) port 1 - FIC A (Port 1 ... for example)
     Chassis 2 IO Module 1 (left) port 1 - FIC A (Port 2 ... for example)
     Chassis 1 IO Module 2 (right) port 1 - FIC B (Port 1 ... for example)
     Chassis 2 IO Module 2 (right) port 1 - FIC B (Port 1 ... for example)
Cheers
Nuno Ferreira

Similar Messages

OSB cluster setup with SSL

Hi,
Could any one help me here..
Cluster setup for OSB with SSL enable
1) Admin and 2 Managed server are running on same host
2) cluster domain created in development mode
2) While starting second managed server getting below error..
<Oct 4, 2010 8:04:58 AM PDT> <Error> <ClusterTimer> <BEA-000000> <Cannot contact Admin server. Therefore constructing the Cluster Authority Current time with
the time skew 0
java.rmi.RemoteException: ClusterTimerAuthority error; nested exception is:
javax.naming.CommunicationException [Root exception is java.net.ConnectException: t3s://hostname:SSLport: Destination unreachable; nest
ed exception is:
java.io.IOException: Stream closed.; No available router to destination]
at com.bea.wli.sb.init.RemoteClusterTimerAuthority.getClusterTimerAuthorityCurrentTime(RemoteClusterTimerAuthority.java:38)
at com.bea.wli.timer.ClusterTimerService.clusterTimeAuthorityCurrentTimeMillis(ClusterTimerService.java:177)
at com.bea.wli.timer.ClusterTimerService.initialize(ClusterTimerService.java:88)
at com.bea.wli.sb.init.FrameworkStarter._preStart(FrameworkStarter.java:221)
at com.bea.wli.sb.init.FrameworkStarter.access$000(FrameworkStarter.java:79)
Truncated. see log file for complete stacktrace
Thanks,
Sushma.

Even I faced the same issue..but eventually the problem got resolved with below resolution:
Resolution: The managed server was not able to connect to t3s://hostname:sslport. The SSL configuration on Adminserver was wrong. After correcting SSL setting on Adminserver..i was able to resolve this error.

Verify HA Cluster setup in Oracle Virtual machine thru GUI

Hi All,
I have setup 3 node cluster with HA[Oracle VM] enabled.When i checked in the GUI it
shows "High availability Infrastructure works well" like that.
For testing HA, I poweroff one of the Oracle Virtual server node from the Oracle Virtual manager browser.
Then if i checked the "Server Pools > Edit Server Pool ".check for "High Availability Infrastructure" browser keep going.
Also one more thing is the Powered off machine automatically after some time it is coming up.Is
it right ?
After powered off if i run the below command in ovmmgr it shows 2 nodes are up.Same o/p
how can i get in GUI when i clicked High Availability Infrastructure in "Server Pools > Edit Server Pool "
[root@xxxovmsr1 ~]# /usr/lib/ovs/ovs-cluster-check master alter-fstab
O2CB cluster ocfs2 already online
2
Cluster setup complete.
Link hangs quit sometime.Not allowing new connections also ..What to do ?
Basic question
How to verify the cluster setup from the GUI when i click "High Availability Infrastructure" check button?
Some one can help me in this please ?
Regards,
JK
Edited by: user2861234 on Jun 29, 2009 8:11 AM

Hey,
After powered off one of VM server if i run the below command in ovmmgr it shows 2 nodes are up.
Why are you going ovmmgr ?
Are your checking it only thru gui part of the manager or also thru thr command line.
Oracle vm server - are hosts for your virtual systems your create on them.
If you power off the server - from GUI part you can see - it as switched off or unreachable
If your HA feature is enabled for that pool having more than 1 server sharing the /OVS filesystem - then he virtual system which is runnign on 1 of the Oracle VM server - will not die
It will migrate to other available server in the pool - which is whole idea and purpose of HA
I don't understand what exactly you are trying to crosscheck or verify here with these commands
High Availability Infrastructure in "Server Pools > Edit Server Pool "
root@xxxoxxxr1 ~# /usr/lib/ovs/ovs-cluster-check master alter-fstab
O2CB cluster ocfs2 already online
2
Cluster setup complete.
If you have to check whether your HA feature is working - form GUI you can got o server pool - tab -
edit it
and chk the HA - tab
and verify-
it HA feature is working properly - it will give you message accordingly.

Can not connect to a cluster and hyper-v manager(to one of my hosts) but virtual machines are running (hyperv2012r2)

Hello guys
I'm using a cluster hyper-v 2012 r2 with 9 hosts and after 12 hours without energy, it back again. my problem is:
my cluster do not back
all hyper-v hosts are and running hypervisor... all vms running... my problem is the cluster.... quorum do not back online...
i found the problem... the wmi is not working in one of my hosts and in that, i cant connect to this hyperv manager... from anywhere.... the virtual machines are running on this host... but i cant manage them...
i found an article talking about this and now, im running this script.... but nothing hapen
"C:\windows\system32\wbem> mofcomp.exe cluswmi.mof
Microsoft (R) MOF Compiler Version 6.3.9600.16384
Copyright (c) Microsoft Corp. 1997-2006. All rights reserved.
Parsing MOF file: cluswmi.mof
MOF file has been successfully parsed
Storing data in the repository..."
someone can help me?
tnx!

Tnx for the ansywer..
"When you say "my
cluster do not back" do you mean that the cluster is not running?"
yes... i cant connect to the cluster
Is "quorum
do not back online" another way you are saying the cluster is not running?
yes
Are you using disk witness or file share witness? Recommend
to use disk witness as it provides a more robust level of operation.
witness
Did you run the cluster validation wizard? Turn off the disk tests to prevent VMs from
failing over, but run all other tests. What errors/warnings do you get?
When i try to run validation tests, one of my hosts (same that i cant manage vitual machines), fail to join in tests but all others are tested

What is the best way to connect a firewall cluster to a VPC domain

Hi All
Can anyone help me decide what is the best way to connect a firewall cluster to a VDC running in a pair of N7K's which is a VPC domain?
Can I configure a VLAN interface on each VDC and use HSRP? I was planning on presenting one 10GB cable from each VDC to each firewall. Would this work OK? HSRP traffic will go across the VPC peer link correct?
thanks all

No, but the one caveat is vpc orphan ports. If the vpc link between the nexus switches fails for any reason, all the vpc ports on the vpc secondary switch will be forced down. So it's recommended to connect single port devices to the primary vpc switch so the connections stay up. But if you're ok with that, then I don't see any problems.
You have a few options, one would be to run a separate link between your nexus switches for non-vpc vlans. These vlans would not be allowed over the vpc peer-link, or forwarded out vpc's.
See here page 49 :
http://www.cisco.com/c/dam/en/us/td/docs/switches/datacenter/sw/design/vpc_design/vpc_best_practices_design_guide.pdf

Wet610n connecting to netgear WDS setup

Hey,
Been trying to get a wet610n to connect to the following setup:
netgear wndr3700 router: the router has bridging and repeating enabled with the mac addresses of the netgear wn802t repeater and the linksys wet610n bridge in its table.
The netgear wn802t will allow wireless client connections just fine, but we cannot get the wet610n to connect. We could get it to connect to the wndr3700 router with the bridging and routing funciton disabled, but then the wn802t will not connect.
I've come to the conclusion that the linksys will not play well with the netgear devices in the mode that they need to be in for repeating.
Any thoughts?
Thanks

If you set the Netgear Router in bridge mode then,the WET will not work because two bridge can not communicate wirelessly;

Java Application connectivity to Coherence cluster

I am looking around the options for Java application connecting to Coherence cluster..
These are the options I seem to have
1. Coherence Extend -
2. Make Java application as a cluster member with no storage
what are the pros/cons?
Thanks

Hi,
Here are my thought off the top of my head
*1. Use Extend if...*
<li> Your application is short lived - i.e. does not run for very long (minutes)
<li> Your application is not located close (in network terms) to the rest of the cluster
<li> Your client uses a different version of Coherence to the cluster (there are caveats around this)
<li> The client application is being developed by another team and its release cycle is not under your control
Pros:
Good for short lived applications
Good for remotely located clients
Can be written in other languages than Java
Cons:
Extend will perform slower than a cluster member as every request goes via the proxy server and then to the rest of the cluster.
If using an invocation service where you want to run invocables across the cluster it needs a two step process.
A badly written Extend client can perform requests which could take out the proxy server
*2. Use a cluster member if...*
<li> Your application runs for a long time - i.e. it is something like a web or application server
<li> Your application is well behaved - i.e. does not have big GC pauses. A very badly behaved cluster member can destabilize the rest of the cluster
<li> Your application is located close (in network terms) to the rest of the cluster
Pros:
Faster than an Extend client
Cons:
Needs to be stable
Needs to be located close to the rest of the cluster
Hope that helps as a start
JK

Why do we use reverse proxy for Oracle RAC Cluster setup

Hello All,
I got this question lately.. "why do we use reverse proxy for Oracle RAC Cluster setup". I know we use the reverse proxy at Middleware level for multiple security reasons.
Thanks..

"why do we use reverse proxy for Oracle RAC Cluster setup".
I wouldn't. I wouldn't use a proxy of any sort for the Cluster Interconnect for sure.
Cheers,
Brian

Libproxy sp12 behind a firewaill fails to connect to weblogic cluster

Two WLS 5.1 SP11 clustering, iPlanet4.1 w/wlproxy plugin, Solaris Box. There is a firewall in between the iPlanet WebServer and the weblogic cluster. The NSAPI proxy plugin ( libproxy.so )connects successfully to one of the weblogic cluster for the first time using the NAT IPs of the cluster m/c maintained in the obj.conf. The Response contains the actual IP of the weblogic cluster m/c which gets updated in the proxy and uses that to connect to the cluster and firewall blocks that as it is not the NAT IP. But this works fine in libproxy of weblogic 6.1 unfortunately this has a problem when the data that gets passed is more than 1000 chars.

shouldn't the connect string look like jdbc:oracle:thin:@MyIP:1521:MySID ?

BAM Cluster setup - issue with whole server migration configuration

HI,
We are working on a SOA , BAM , BPM Cluster setup on OEL VMs
With Admin server running on Host1 and Soa_server1, BAM_server1 , Host2 - Soa_server2 , Bam_server2.
We were successfully able to setup clustering for Soa without using the webTier utilities in our environment, Just by adding entries in the /etc/hosts file.
Followed the document - except for web tier setup
http://www.namredlek.nl/orasoa/InstallSOASuite11gOnCluster_v1.pdf
1)     Installed Weblogic 11g.
2)     Soa 11g.
3)     Domain creation
4) Configured Node manager
5) Soa Clustering completed successfully.
6) Domain up and running with SOA_Cluster up and running.
7)     Extended the domain with BAM
The issue revolves when we are doing teh Migration configurations from the console for BAM and we restart BAM server we get an error -
Apr 14, 2011 2:30:35 AM weblogic.nodemanager.server.Handler handleStart
WARNING: Exception while starting server 'bam_server1'
java.io.IOException: Command '/oracle/Shared_MWHome/user_projects/domains/SOA_Domain/bin/server_migration/wlsifconfig.sh -addif eth0 192.168.1.6 255.255.255.0 ' returned an unsuccessful exit code '1'. Check NM logs for script output.
and Bam server goes to Failed Not Restartable state.

The error is clearly out there
java.io.IOException: Command '/oracle/Shared_MWHome/user_projects/domains/SOA_Domain/bin/server_migration/wlsifconfig.sh -addif eth0 192.168.1.6 255.255.255.0 ' returned an unsuccessful exit code '1'. Check NM logs for script output.
This error occurs due to a failure in adding the virtual IP of BAM server to the network interface you have specified in nodemanager.properties.
You need to check whether you have specified the correct interface and subnet mask in node manager.properties as below :
The fourth step is to edit the Node Manager's properties file. This file is called nodemanager.properties
and is located in the ORACLE_BASE/product/osbmw/wlserver_10.3/common/nodemanager directory.
For server migration to work properly, you need to add the properties listed below:
Interface=eth0
NetMask=255.255.255.0
UseMACBroadcast=true
· Interface
This property specifies the interface name for the floating IP (for example, eth0).
· NetMask
This property specifies the net mask for the interface for the floating IP. The net mask should the
same as the net mask on the interface; 255.255.255.0 is used as an example in this document.
· UseMACBroadcast
This property specifies whether or not to use a node's MAC address when sending ARP packets,
that is, whether or not to use the -b flag in the arping command.

SOAP ports with cluster setup and BES?

Okay, our cluster setup in NetWare (yes, still NetWare cluster) with Groupwise 7.0.3 we have unique TCP ports for the client/server (like 1677-something). Even though each clustered node has a unique IP and is running in protected memory.
That being said:
We'd like to enable SOAP for use with BES and GWArchive.
The issue/question:
Do either of these figure out the SOAP port for each PO? Or must the SOAP port be the same among all the PO's for BES to work properly?
If the SOAP port must be the same (I'm not the BES admin, so I don't know if the setup lets you specify just one port or if you can defined them per PO), can you run multiple POA on the same physical host (keep in mind, this is clustered with protected memory) and use the same TCP port # for SOAP?

kjhurni wrote:
> Okay, our cluster setup in NetWare (yes, still NetWare cluster) with
> Groupwise 7.0.3 we have unique TCP ports for the client/server (like
> 1677-something). Even though each clustered node has a unique IP and is
> running in protected memory.
>
> That being said:
>
> We'd like to enable SOAP for use with BES and GWArchive.
>
> The issue/question:
>
> Do either of these figure out the SOAP port for each PO? Or must the
> SOAP port be the same among all the PO's for BES to work properly?
>
> If the SOAP port must be the same (I'm not the BES admin, so I don't
> know if the setup lets you specify just one port or if you can defined
> them per PO), can you run multiple POA on the same physical host (keep
> in mind, this is clustered with protected memory) and use the same TCP
> port # for SOAP?
>
>
I haven't used BES, but in general there's no auto-discover method for
SOAP ports, so yes they'd probably have to be the same port.

Connecting to the cluster

Hi,
          I have a cluster running on a multihomed Solaris Server. It is made of two
          servers running on IP1 and IP2 (both using port 7001). It says in the
          documentation that I can put in the IPs of the two servers separated by a
          comma for the cluster address. So the address that I put in was
          "IP1:7001,IP2:7001" instead of a DNS name.
          How do I connect to the cluster. If I connect to IP1:7001 then I am
          connecting to the server and not the cluster, correct?
          I would appreciate any help,
          Mohammad


Sorry for the delay, I have been extremely busy...
          There are several alternatives -- DNS round-robin, hardware load balancers
          (e.g., Cisco LocalDirector), etc.
          Mica Cooper wrote:
          > Robert,
          >
          > WLBS is Windows Load Balancing Service. It was designed to only use 1 IP per
          > box. It comes on the server editions. Do you know of something else? Linux
          > or Solaris maybe?
          >
          > Mica
          >
          > "Robert Patrick" <[email protected]> wrote in message
          > news:[email protected]...
          > > Hi,
          > >
          > > I'm not sure what WLBS is but I guess my question is does it really need
          > to know
          > > that the IP addresses happen to point to the same machine? What happens
          > if you
          > > just tell it that there are twice as many machines as there really are?
          > >
          > > Robert
          > >
          > > Mica Cooper wrote:
          > >
          > > > Currently,
          > > > I am using WLBS to do the load balancing and it will only handle one IP
          > per
          > > > box. Do you know of any other software to do this?
          > > > Mica
          > > >
          > > > "Robert Patrick" <[email protected]> wrote in message
          > > > news:[email protected]...
          > > > > The real answer here is that the "cluster" is virtual -- it only
          > exists
          > > > because
          > > > > there are one or more server instances running. Regardless of how you
          > > > attain
          > > > > it, you are going to connect to one or more of the server instances.
          > > > There are
          > > > > several ways to do this as has been previously discussed (e.g., DNS
          > alias
          > > > that
          > > > > does DNS round-robin-style load-balancing, a comma-separated list of
          > IP
          > > > > addresses, etc.).
          > > > >
          > > > > Mica, as for your question, you need to teach your "load-balancing IP"
          > > > about all
          > > > > of the IP addresses for the servers in the cluster and not just "one
          > IP
          > > > per
          > > > > physical machine".
          > > > >
          > > > > Hope this helps,
          > > > > Robert
          > > > >
          > > > > Mica Cooper wrote:
          > > > >
          > > > > > Tao,
          > > > > >
          > > > > > Thats not what he is asking.
          > > > > >
          > > > > > He's asking how to call MyServerClusterName instead of trying to
          > call
          > > > > > different instances by IP. I would like to know how to do this also.
          > > > > > Currently we are calling a load balancing IP and it proxies to each
          > > > instance
          > > > > > but it only works for 1 instance per box and I need 2 instances per
          > box.
          > > > > >
          > > > > > Mica Cooper
          > > > > >
          > > > > > "Tao Zhang" <[email protected]> wrote in message
          > > > > > news:[email protected]...
          > > > > > > If you don't use DNS, you can write the IP1:7001,IP2:7001 in your
          > > > > > > PROVIDER_URL, and then transfer it to InitialContext.
          > > > > > >
          > > > > > >
          > > > > > > Mohammad Khan <[email protected]> wrote in message
          > > > > > > news:[email protected]...
          > > > > > > > Hi,
          > > > > > > >
          > > > > > > > I have a cluster running on a multihomed Solaris Server. It is
          > made
          > > > of
          > > > > > two
          > > > > > > > servers running on IP1 and IP2 (both using port 7001). It says
          > in
          > > > the
          > > > > > > > documentation that I can put in the IPs of the two servers
          > separated
          > > > by
          > > > > > a
          > > > > > > > comma for the cluster address. So the address that I put in was
          > > > > > > > "IP1:7001,IP2:7001" instead of a DNS name.
          > > > > > > >
          > > > > > > > How do I connect to the cluster. If I connect to IP1:7001 then I
          > am
          > > > > > > > connecting to the server and not the cluster, correct?
          > > > > > > >
          > > > > > > > I would appreciate any help,
          > > > > > > > Mohammad
          > > > > > > >
          > > > > > > >
          > > > > > > >
          > > > > > >
          > > > > > >
          > > > >
          > >

How to Uninstall SQL instance on active-passive SQL server , which failed during Cluster Setup (Error-Failed at Validate Active Directory Configuration)

How to Uninstall SQL instance on active-passive SQL server , which failed during Cluster Setup (Error-Failed at Validate Active Directory Configuration)
active-passive SQL server cluster setup failed due to some steps missed in initial cluster setup,
now i have unistall sql instance from nodes,
Your help will higly appriciated.
Regards,
Anish
Asandeen

Hello,
Please refer to the following link about remove a node of SQL Server Failover Cluster Instance:
http://msdn.microsoft.com/en-us/library/ms191545.aspx#Remove
Regards,
Fanny Liu
Fanny Liu
TechNet Community Support

Display logfiles in cluster setup (HTTPS UI)

Hi all
In my standalone setup, I can display and download logfiles in HTTPS-Webgui under "Log Subscriptions" -> "Log Files".
But in a cluster setup, I can't see the column "Log Files" under "Log Subscriptions". How can I access the files?
Thanks!
Christoph

Hello Christoph,
viewing individual logfiles on the GUI is notavailable for appliances in a cluster. You can still access them via your browser by using:
ftp://appliancename.local
as URL, where "appliancename.local" is the hostname or IP. This will prompt you for an username and password, you can use your normal admin credentials. Also possible:
ftp://[email protected]
ftp://admin:[email protected]
If you want to include username (and password)
Hope that helps,
Andreas

Network Extender & Cluster Setup

Does the Ingress unit satisfy the GPS signal requirements for other non-Ingress units in a cluster setup?
We have a large steel warehouse that has no cell coverage in the building. I can provide coverage with 3 Network Extenders. However, my problem is going to be the GPS antenna's. I have 1 window that provides a good GPS signal for the 1 extender I have now. If I purchase 2 additional extenders to complete the AoC will each of them have to lock on to GPS too? Or do the non-ingress units not require GPS. It may be wishful thinking, but our only other option is to cut a hole to access the outside & I really dont think these antenna' can handle the environment.
All units will be on the same network switch.

Ah. And I am guessing (really just confirming) that Open Directory is one piece of functionality which is in OSX Server, but not in the basic Snow Leopard distribution.
Sooo, is it worth digging around to see if any of the linux distributions out there contain Open Directory? I have an old lime (333 MHz) iMac in the basement...

IOM connections in a cluster setup

Similar Messages

Maybe you are looking for