I've installed WebSphere Application Server Network Deployment and need to run three different applications.
For each (Standalone) Application Profile creation I come across these options
Node Name, Server Name & Host name.
When first profile is created the values are
Node name - iamNode01
Server name - server1
Host name - iam
Second profile has following values auto-populated
Node name - iamNode02
Server name - server1
Host name - iam
etc..
If any one could help me understand on Node name and server name part.
There are two types of profile, you can have a single server/standalone profile, or one that is part of a cell and managed by a deployment manager. If you are using standalone profiles I would recommend having one server per profile. If you are using a deployment manager managed profile then you can have multiple servers per profile.
When an ND installation spans more than one machine, each machine has to be a different node. You can have multiple nodes on a single machine too if you like.
If there's a deployment manager in use, the simplest single machine setup is two profiles, one with the deployment manager in it and the other with all the appservers in it.
If there's no deployment manager (a standalone profile), then the simplest thing is one appserver per profile, and each profile will have it's own nodename.
Node name usually identifies logical node, in case of distributed installations (the ones with deployment manager), it helps you identify machine where servers are installed.
In case of standalone profiles, it is just hostname with Node0# appended.
If you want to have multiple standalone profiles on one machine, you can consider creating AdminAgent profile and register standalone profiles there. It will allow you to administer all servers from one console, just selecting specific profile on login.
You could create many servers on single standalone profile, however it is only possible via wsadmin and in general not recommended.
Server name identifies specific JVM, it must be unique on given node.
Maybe an element you are missing here is that Standalone profiles are not coordinated by a deployment manager nor node agent, so they are each just acting as if they are the a singler server on a single node -- hence the apparent duplication/multiple nodes per host.
You wouldn't have this relationship in federated (non-standalone) profiles -- just 1 managed node per operating system instance.
Related
I have several Karaf-Containers in different environments, each with 50+ services deployed, and I´m looking for an easy way to compare the versions in each environment.
Is it possible to retrieve a list of all deployed services including the specific version via the jmx-Interface?
You can use the BundlesMBean with objectName org.apache.karaf:type=bundle,name=root.
The attribute Bundles returns Tabular data that contains among other details Symbolic Name and Version.
Simply try with jconsole. It is easy to connect to a local karaf installation.
I'm trying to do some clustering testing and I am setting up multiple RabbitMQ services on a single Windows machine. I am able to set the environment variables RABBITMQ_NODENAME, RABBITMQ_SERVICENAME, and RABBITMQ_NODE_PORT then run RabbitMQ-Service Install to have a new RabbitMQ service installed under a different name.
My question is regarding the configuration file. Based on what I read on the RabbitMQ site, the configuration file defaults to the %AppData%\RabbitMQ directory.
I'm just having trouble trying to understand how it should be setup so I can have 3 instances of the service running with their own configuration.
Do I run the installation under a different local or domain account so it gets placed under a different %AppData%\RabbitMQ directory or can I add a directive to the service to look in a particular directory for the configuration file for that particular service?
Also, how does RABBITMQ_BASE come into play? Is that only for data and log files or does that also apply to the configuration file? I'm not sure if once I have the service setup with BASE defined as a specific path I can place a new rabbitmq.config under the root of that path.
Please confirm and provide any additional assistance. Thank you in advance!
For now I'm testing on Windows but I plan on converting to linux once I have this all working correctly and understood. Unfortunately, I've inherited the current environment and it's already installed and running using Windows servers. They just wanted me to setup clustering for it so I'm trying to simulate the cluster on my workstation.
Nevermind, I found out what I needed. The environment variable RABBITMQ_CONFIG_FILE can be used to override the location of the default config file.
http://www.rabbitmq.com/relocate.html
You can run multiple RabbitMQ instances on 1 machine without clustering. You just need to change the ports and the node name in rabbitmq-defaults, rabbitmq-env and config files. If you want them as a service you can just create them from the already configured instances.
HERE is a detailed guide on how to do that. It's pretty easy and straightforward.
THE MISSION:
I have a development environment running on an Amazon AWS EC2 virtual server which i want to have tested by third parties.
THE PROBLEM:
I do NOT trust the companies who will test it not to sabotage environment and / or steal code. Therefore, i don't want them to know URL's, permanent IP's or even to access the web pages, which they could eventually use a crawler to find.
My environment includes web applications and socket servers. I do NOT want to expose the web applications, while giving access only to socket servers.
THE CONCEPT:
I have opted to use a secondary, impermanent Elastic IP pointing to the environment. this IP will be destroyed after 1 or 2 days, after basic tests have run. Subject to change (depending on suggestions from this thread).
THE QUESTION:
Can i create a secondary Elastic IP instance that allows access only to ports 5000-5100? If so, how?
THE ALTERNATIVE: In case this is not the most efficient procedure, what alternative would you propose?
MY SOLUTIONS: followed FAQ Launching Instance From Backup
create snapshot
create image from snapshot (snapshot menu - create image tag)
instances - launch instance
choose image created from snapshot as your root volume
edit security groups (opened port range for sockets only, no web)
deleted all web code from this instance
after 2 days, will delete instance
followed Create Image From, Instance
select (exclusively) running instance you wish to mirror
right click on selected instance
choose create image from dropdown
to 7. same as above
this second solution seems to be more stable (especially re: status check and connectivity issues).
any better solutions? thanx!
I've been watching some videos from the build conference re: Inside Windows Azure etc.
My take away from one of them was that unless I loaded in a preconfigured VHD into a virtual machine role, I would lose any system settings that I might have made should the instance be brought down or recycled.
So for instance, I have a single account with 2 Web Roles running multiple (small) websites. To make that happen I had to adjust the settings in the Hosts file. I know my websites will be carried over in the event of failure because they are defined in the ServiceConfiguration.csfg but will my hosts file settings also carry over to a fresh instance in the event of a failure?
i.e. how deep/comprehensive is my "template" with a web role?
The hosts file will be reconstructed on any full redeployment or reimage.
In general, you should avoid relying on changes to any file that is created by the operating system. If your application is migrated to another server it will be running on a new virtual machine with its own new copy of Windows, and so the changes will suddenly appear to have vanished.
The same will happen if you perform a deployment to the Azure "staging" environment and then perform a "swap VIP": the "staging" environment will not have the changes made to the operating system file.
Microsoft intentionally don't publish inner details of what Azure images look like as they will most likely change in future, but currently
drive C: holds the boot partition, logs, temporary data and is small
drive D: holds a Windows image
drive E: or F: holds your application
On a full deployment, or a re-image, you receive a new virtual machine so all three drives are re-created. On an upgrade, the virtual machine continues to run but the load balancer migrates traffic away while the new version of the application is deployed to drive F:. Drive E: is then removed.
So, answering your question directly, the "template" is for drive E: -- anything else is subject to change without your knowledge, and can't be relied on.
Azure provides Startup Scripts so that you can make configuration changes on instance startup. Often these are used to install additional OS components or make IIS-configuration changes (like disabling idle timeouts).
See http://blogs.msdn.com/b/lucascan/archive/2011/09/30/using-a-windows-azure-startup-script-to-prevent-your-site-from-being-shutdown.aspx for an example.
The existing answers are technically correct and answer the question, but hosting multiple web sites in a single web role doesn't require editing the hosts file at all. Just define multiple web sites (with different host headers) in your ServiceDefinition.csdef. See http://msdn.microsoft.com/en-us/library/gg433110.aspx
I'm connected to one of my WebRole instances. But what can I do now? I startet the Server Manager and navigated to Roles -> WebServer -> RD00123..., but what I see seems not to be the active instance. When I click on "Browse WebSite" I don't see the deployed site even I can't see or edit for instance the configured rewrite rules (defined in the web.config).
It's also funny that the inetpub is on systemdrive D (d:\inetpub\wwwroot) but the deployed (editable) files are on E (e:\approot). I'm wondering how this comes togehter.
Ok the things I want to know are:
How can I launch the unique instance version of the web role (I can't find an IP or port)?
How can I assume temporary debug changes in the web.config?
How can I restore the deployed version of the instance?
1) I don't believe you can. This is managed by the Azure Fabric controller. The most you can do is force an individual instance (the one you've RDP'd into) to reboot/restart.
2) Yes. But they won't be reverted unless the VM is recycled, so be prepared to back them out yourself. This is a good use of RDP
3) Not directly. Again, this is controlled by the Azure Fabric controller. But you could put the original bits you had back into the VM manually.
I'm not sure I understand the questions regarding server manager, but I can answer the question about where things live. The "base" Windows Azure VM image is the root drive where you see INetPub at. The application is deployed to a seperate VHD (thus its presence on another drive). If you request local storage, that will come from yet another virtual drive. This is just how the Azure Fabric manages the application deployments. Properly constructed, the roles of course won't care where things live.