Our application accepts incoming requests (REST) and satisfies them using another of our services. When this application boots, it requests a number of connections from that other service - our goal is to maintain a 1:1, thread-to-connection, ratio (I won't get into why, its just that way).
Naturally, we want to define the number of threads/connections in one place, my thought is for the webapp to discover the "maxthreads" value as configured in server.xml -- without having to navigate to and parse the server.xml file. Is this possible?
Thanks
I'll answer my own question - JMX / MBeans.
Related
I have multiple web applications running under a single Tomcat container. Since they all run under a single Tomcat connector (as defined in the server.xml file), attributes such as maxConnections and maxThreads govern the container as a whole. As a result it is possible for a single application to consume all available Tomcat threads, starving the other applications of threads and making them unresponsive. I would like to be able to define the maximum http threads on a per context basis so that this is no longer possible.
Here's what I've tried so far:
Create a custom filter in the application that keeps track of the current thread count and limits additional connections. (Got the filter here: How to set limit to the number of concurrent request in servlet?). I'm not sure I like this solution, as it isn't as full-featured (support for attributes such as acceptCount, maxConnections, maxThreads, and minSpareThreads) as Tomcat provides by default to the container; and adding in the features feels like I am attempting to build what already exists in Tomcat.
Create a separate Tomcat connector in the server.xml file for each context. This has a few issues. For one, each connector requires a separate port; this means I'll have to account for this in my apache config. Secondly, I plan to add more webapps regularly; this means a config change followed by a tomcat restart, which is disruptive to clients.
Has anyone else encountered something like this? I feel like there should be a "Tomcat supported" workflow to accomplish what I'm after.
I'm going to post an answer that was provided to me from the Tomcat user group: http://tomcat.apache.org/tomcat-9.0-doc/config/valve.html#Semaphore_Valve (The Semaphore Valve is not Tomcat 9 specific, but was actually introduced in Tomcat 6). I experimented with this concept, and I found the following practical applications:
(Untested) The Semaphore Valve should be able to be nested within the Host element in the server.xml file.
(Tested) A [context-name].xml file can be placed inside [tomcat-home]/conf/Catalina/localhost with the valve nested within the Context element.
This is not necessarily the solution that I am going with, as more testing will need to be performed. However, I thought I'd add this as it is a potential answer to the problem.
Update:
As a recap, the SemaphoreValve was an option that was recommended to me through the Tomcat user mailing list as a solution to the issue that I described above. It turns out it was easier to implement than I anticipated. Adding the following to context.xml in the Tomcat/conf directory did the trick:
<Valve className="org.apache.catalina.valves.SemaphoreValve"
concurrency="10"
fairness="true" />
Thanks to Mark Thomas from the Apache group for supplying the solution.
I have an application with name test.war.
Because of Apache installed on my server I have to use another port number for Tomcat applications so after deployment this application available at domain.com:8080/test/.
I decided to create a subdomain in order to remove that ugly 8080 from url, so I setted up the server like described here. So now test.domain.com reffers to domain.com:8080/test/.
Everything seems fine except of one problem - because my application is not ROOT.war and I am using spring:url function every url in the application is translated to /test/bla-bla. So I removed that function. But still have a problem with spring security because it still translates an urls relative to app name i.e. /test/bla-bla.
How can I solve that problem?
Thank you
UPD: I don't want to deploy it as a ROOT application because I have two or three such applications and I wanted to create a subdomain for each one of them
Spring Security doesn't "translate" URLs. In fact this isn't specific to Spring Security. You'll run into similar issues with any application which does redirects. If you want to hide the context paths of applications which are behind a proxy, then you need to rewrite the URLs at the proxy.
I'd recommend you read the Tomcat Generic Proxy Howto and the section on URL rewriting in particular, as it specifically addresses this issue.
We have developed a JEE5 web application (WAR) and running it in production under WebLogic 11g (10.3.5).
Now the same application should be deployed as separate applications for different customers (different URLs, different data) on the same WebLogic.
I managed the first part by setting different context roots after deployment for each of them.
But I have yet to make them use different datasources - and since I want to avoid customer specific builds, the persistence.xml is the same for all applications, thus also the persistence unit name.
What is the best setup for this scenario? Am I forced making separate builds and by that different WARs or do I have to separate Managed Servers or Domains wihtin the server or is there a better way to solve it?
I know this thread is very old,but replying so that it may help someone with the same question stumbling on this thread.
The latest weblogic 12.2.1 comes with Multi-tenancy(add-on I guess) which can let you run same applications in a single domain.
Edit: Weblogic 12.2.1 introduced concept called Partitions. Partitions are both config and run-time subdivision of a weblogic Domain. In a single weblogic domain you can create multiple partitions. Each partition will have one or more resource groups. Resource groups are the logical grouping of weblogic resorces like data sources,jms,Java EE apps ,etc. For example to achieve what the original posts asked for , we create a Resource Group template with the web-application and the datasource as the resources. In the Data source configuration we can provide a place holder variable instead of actual URL as DB URL. Then we can create two partitions that refers to this Resource Group Template(Each partition will now have a separate web application and data source) . Each partition will override the DB URL property there by creating two data sources with same JNDI name.In each Partition we create virtual host/port so that the client can use that to access the application running in the respective partitions.
A better and more detailed information on this can be found in https://blogs.oracle.com/WebLogicServer/entry/domain_partitions_for_multi_tenancy
ServletContextListener.contextInitialized can look at the ServletContext and figure out which deployment is which
in web.xml, define a servlet context listener:
<listener>
<listener-class>com.path.YourServletContextListener</listener-class>
</listener>
and then in YourServletContextListener.java, add a contextInitialized method like this:
public void contextInitialized(ServletContextEvent sce)
{
ServletContext sc = sce.getServletContext();
String name = sc.getContextPath();
...
}
my thought is that you can use that name to select from multiple data sources that you have configured. depending on how you've been deployed, you'll make a different database connection and have the correct application's data.
It seems to me from what I saw in the Oracle documentation, that having several domains is the only way to separate data sources with the same persistence unit name - which is bad, since this basically means running two WLS in parallel.
For this reason I decided to go with building individual WAR files (which I tried to avoid initially), to include customer-specific persistence.xml files and specifying customer-specific datasources in the WLS.
I want to host a tomcat server, which will host 3 web applications. I want to get the total number of requests processed by my server(for any of the 3 web applications). Is there any log from where I can check the total number of requests processed by my server(including HTTP 404 requests)
Note : I can calculate the total requests processed by each individual application and get the total number of requests processed by my server, but the applications deployed in tomcat are third-party and i cant do any changes in it.
Basically, I am building a monitoring application for the tomcat server and I have to provide the total requests served in the application,
Also, My first thought was to over-ride the HTTPServletRequest class constructor in servlet-api.jar and put a static counter. since every request is mapped to HTTPServletRequest object, I guess it will do the job. But is it a good idea to over-ride the HTTPServletRequest or is there any existing solution available for this?
If you do choose to override the HTTPServletRequest class do not just add a static counter, this will cause your counter to reset itself each time the Server goes down/jvm reloads.
I think it might be a better option to either increment it from a database or save the value in a file each time. This way you do not loose your count even if something happens to the server and you have to restart it.
This is all assuming that there isn't already a Apache extension that already dose this and you do want to tinker with the HTTPServletRequest class.
You can use a special filter called Request Dumper Filter for this purpose.
According to official Tomcat Documentation:
The following entries in a web application's web.xml would enable the Request Dumper
filter for all requests for that web application. If the entries were added to
CATALINA_BASE/conf/web.xml, the Request Dumper Filter would be enabled for all web
applications.
My old way using mod_jk in apache and configure virtual host in tomcat
In the JSP file, I refer to CSS as below
/<%=request.getContextPath()%>/css/styles.css
while the home link is set to
/<%=request.getContextPath()%>/
so this worked fine when I use mod_jk in apache to work with tomcat using ajp;
When I try to configure reverse proxy as below
ProxyPass / http://localhost:800/mywebapp
ProxyPassReverse / http://localhost:800/mywebapp
the home page can be retrieved fine but the css request becomes
http://mydomain.com/mywebapp/mywebapp/css/style.css
so the css file can not be retrieved correctly;
I think one possible way is to always use relative path like ./style.css or ../style.css
a. since header/footer are shared, and the home page is in a different level with detail page, it's inconvenient to use relative path because they're at a different level
b. still, I think the home link will have to be /<%=request.getContextPath()%>/
so I wonder what's the way to set contextroot fine in java web and also work fine with reverse proxy?
thx a lot
As I know your application server (Tomcat) isn't able to be aware of a reverse proxy presence. Generally speaking, it can be contacted through any number of reverse proxies or directly by browsers. A network configuration is usually used to limit this, not HTTP or Java.
So, you must accurately rely on relative URLs to make your application work well.
When I have to deal with reverse proxy presence (almost always due to SSO architectures), I embed a "junction" configuration string item (the portion of the URL used in the proxy to map the application) and use it in the only places where I need to build an absolute URL.