i have 2 machines on which same version of glassfish is installed say A and B. they are nix based. on both the machines glassfish is hosting same applications. glasfish B is facing some issue which we are not able to sort. glassfish A and applications on it are working perfectly. can i take a backup-domain on glassfish A and restore it on glassfish B. ? in other words do domain configuration contains machine specific information or anything that might make this thing not possible ?
Yes. You should be able to do that. We went to lengths to make the configuration not machine-specific.
I recommend that you first backup GF "B" just to be safe.
If you have GF server instances (not domain) running on another machine, they have a config file that points back to the domain's host. You would need to edit that file if the domain host changed.
Related
I've created multiple managed servers in Weblogic. Each managed server has same IP(IP of weblogic server) and unique port.
I've installed ear and war files in these managed servers and all of them has their own context-root.
For example I have ear file A with context-root A on server A and a war file B with context-root B on server B.
Whenever I browse IP:PORT(A) which has set on server A, this application needs some files on war file B which has deployed in server B. So A can not find its resources and I get error.
I should mention that I put them in a cluster too but nothing changed.
How can handle this matter?
The cluster won't help you in this way because each web app is isolated from others in its own web container. If you need to share data or resources between web apps you can use a shared file system or a database for instance.
For this purpose I've used Oracle Http Server (OHS) plugin. after hours of searching in internet I've found that Weblogic application server does not implement anything to handle this issue.
https://docs.oracle.com/middleware/1213/webtier/
I am trying to setup a weblogic cluster running ATG Commerce. I have installed weblogic on machine A, with the following configuration.
Weblogic admin server
Managed servers: Production_A, Production_B, Publishing_A and Publishing_B that
shall run on machine A and B accordingly.
Do I have to install weblogic on machine B also?
(While installation of ATG commerce it ask which is our application server. I mean to say the production and publishing server shall run inside the weblogic server of machine A).
Secondly, do I have to actually specify the managed servers in weblogic while installing it on machine A. I mean to say while installing ATG commerce in machine B, during CIM configuration, I specify the weblogic of machine A, and I create production_B and publishing_B to run inside A's weblogic.
I am confused wheater I'm doing it right.
The managed servers production_B and publishing_B appear in A's weblogic After I do CIM configuration on B. The I set up machines(A and B in A's weblogic), add servers to machines, Add servers to cluster. Everything is happening. But when I try to run B's servers from A's weblogic, the servers do not start saying nodemanager is unreachable. In machine A, I have set the node manager for machine B to Machine A(I need to ask like will the node manager for machine B's server? Or it will run in A's weblogic node manager(production_A and Publishing_A runs with this node manager)) machine A is reachable from machine B(I can open weblogic console of Machine A on Machine B). Am I missing on anything?
Can Anyone direct me to a reference/blog for weblogic clustering in ATG Commerce
Firstly, you will need to install WebLogic on every physical machine on which you intend to deploy your EAR. It is the servlet container that you'll be using so without it (and it's obvious dependencies like JAVA) you won't be able to run your deployments on that machine.
As far as your ATG instances are concerned, I would do it as follow:
Create Commerce A and Publishing A on Server A (using CIM). Something that you are missing though is you have no LockManager configured. You'll require at least one of these to maintain your locks across your Commerce Instances and possibly another to do the same across your publishing instances (I've never deployed a clustered publishing environment, only ever one BCC per environment).
Having done the configuration on Server A, manually copy the ATG-Data/servers folder (or /servers) folder from Server A to Server B. Because you don't install ATG on every machine, in fact you don't need ATG to be deployed if you create your EAR in standalone mode, I would recommend you setup an ATG-Data folder on both Server A and Server B and deploy your configs in there.
Now, having copied the servers folder you will need to manually edit the following files:
Configuration.properties
This probably contains references to Server A that you want to update.
The Ports on Server B can be the same (per instance) as it is on Server A
Update the otherLockServer property in the LockManager instance (if you created it) for both Server A and Server B to reference the 'other lock server'
Update the ClientLockManager to point at both LockManagers
From a WebLogic point of view you need to create the servers on each instance as well. Even within a WebLogic Cluster (which is licensed separately with Oracle and not included in your ATG License) you need an AdminServer per WebLogic. I believe the NodeManager configuration would be different but I've not set this up in a WebLogic cluster, yet.
I'm using a Debian VPS on DreamHost and wanted to install a feature-rich customer issue-tracking system (not for software development like Bugzilla). OTRS made my shortlist and I followed the Installation Instructions through the "Web Server Configuration" step (/etc/init.d/apache2 restart), but the restart step reported failure. Nevertheless pgrep apache showed it was running. In fact, it turned out that although www.mysite/otrs/installer.pl was running, my regular website showed a page claiming it had no content (but when I looked in the website's folder, its content was fine, just not being served).
DreamHost Support was very helpful, but explained that they don't use the standard Debian Apache server for hosting websites and instead use their own. Specifically, the Debian server is in /etc/apache2, but the DreamHost server is in /dh/apache2. DreamHost Support determined that the OTRS installation instructions were configuring the usual Debian Apache location which somehow prioritized that server instead of the DreamHost server. They tried moving the otrs.conf file into /dh/apache2, but though the regular website was working again, the OTRS page wasn't.
Has anybody had success installing OTRS on a DreamHost VPS?
I've consulted one of our admins on this, and these are our suggestions:
You will either need to:
Adapt DreamHost's Apache build to incorporate the OTRS modifications
Get Debian Apache up and running
Both options will require an admin user and some knowledge of Linux command line and Apache management tools. You will also need to set your VPS to UNMANAGED, which means that any changes in the DH Web Panel to any of your domains will have no effect whatsoever. Just make sure the DNS records for any domains are pointing to your server. You will also need to be able to manage your own Apache configurations.
NOTE: This will also essentially mean that DreamHost support cannot and will not troubleshoot your domains. Unmanaged means unsupported in any way!
There are a few core differences between DreamHost's apache2 configuration and the default Debian build. The first issue I observe is that DreamHost's configuration does not allow for extra configuration files to be loaded in the manner that the OTRS documentation suggests. This means if choosing option 1, you will need to manually insert the OTRS directives into DreamHost's configuration files, which may prove difficult.
I would recommend moving or otherwise disabling the /dh folder entirely after setting your VPS to unmanaged. This will not allow DH-default Apache to start when the VPS starts. You may also need to remove the DH Apache startup script in /etc/rc3.d/S02httpd2 and the actual script at /etc/init.d/httpd2.
Once you have your own version of Apache running successfully, you might consider copying the VirtualHosts that were previously at /dh/apache2/apache2-ps/etc/httpd.conf into your own domain configuration files in your conf.d directory, or you can shuffle your website files around and configure your Apache to your desire.
Once you've got your own flavor of Apache running, you should be able to implement the OTRS instructions per their wiki. :)
I'm migrating a lot of websites from Resin 3 to Tomcat 7 (centos 4/apache 2.20) and I'm struggling to determine what type of configuration matches my requirements. In particular:
proxy_ajp vs mod_jk vs mod_proxy for passing requests to Tomcat/Railo
automating deployment of new sites
putting WEB-INF outside the site roots (to simplify cloning sites)
using apache itk with tomcat so each vhost runs as a different user and process
having a single shared railo server administrator config
support for SES URLs with no extension (ie: /path/to/page)
SSL support required
I've read a lot of howtos already but most are out of date or provide conflicting advice. I would like to see some examples from people who run many railo vhosts and deploy them automatically or programmatically. In general I'd prefer efficiency/speed over simplicity as I want to get the most out of limited resources.
I could have asked these questions separately but I want to be sure any answers take into account all the above factors (assuming the requirements are actually compatible).
firstly, check out the vivotech installers - they are a hosting company, so use their installers as your base, they are flawless. (it uses tomcat)
railo 3.3 makes it a lot easier to deploy contexts from admin, so scripting this shouldn't be that hard.
web-inf should be automatically put into a site when it is defined in tomcat
if you give each user a new context-root, then they will have their own admin
every webserver (apache/iis2k8/even tomcat) supports url-rewrite
everything supports ssl
you might also want to look at how you're going to tune your jvm's for this senario, then do some load testing to see how they fare.
drop an email to sean corfield, google railo and his name and you'll get his email.
We are running Weblogic 7sp6. We have a working single node cluster with an Admin and two Managed servers. We are re-creating a 2nd standalone cluster on a 2nd server. We reinstalled Weblogic and have copied over all the configuration files to make thing. Its the same on both clusters. We changed all the references to IP and hostnames. We have used this method before without problems.
In the current case I can startup the Admin which listens on port 7001,7002. But when I try and startup either of the Managed servers it tells me that myserver1/2 is already up. (Managed Servers). I confirmed that myserver is configured to use port 7012,7013 and I cannot find any port conflicts especially because these exact ports worked on the first cluster. Any ideas of what else to look at? I have logged in the admin console and can see the ports are all unique. Thanks
The current version of WebLogic is 10.3. I'd strongly urge you to upgrade your WebLogic as soon as possible, especially if you're still using the version of JDK that it was certified for. If you're running JDK 1.4, you're crazy.