Apache or OHS

Which is better for weblogic 12c, the oracle http server (OHS) or core apache? I know that OHS is a customized version of apache for use with oracle applications. but can ohs handle many requests as apache, and is it a good practice to use OHS with weblogic 12c for production environments, or does it have some issues?

Hi Kal,
Oracle http server offers extra modules than that of Apache.
1. mod_certheaders
2. mod_dms
3. mod_onsint
4 mod_oradav
5 mod_ossl
6 mod_osso
7 mod_plsql
8 mod_wl_ohs
Also in case of any issues or bugs that you encounter with this webserver, Oracle will be able to get a fix .
Thanks,
Sharmela

Similar Messages

Passing Apache (or OHS) environment variables to WebLogic

Hi,
AJP is the protocol used to proxy a Tomcat or Jetty container with Apache HTTP server. Apache environment variables can be passed along to the Java container, instead of using rather insecure HTTP headers.
Is there a way to do the same thing using Apache and mod_weblogic, or using OHS and mod_wl_ohs? From what I've read, these modules are rather pure HTTP proxies. See for instance this post [1].
The idea behind this is to integrate a Shibboleth Service Provider (mod_shib on Apache) with WebLogic, using an Identity Assertion Provider in WebLogic to establish user identities. Shibboleth advises NOT to use headers, and use environment variables instead (see [2]), but it seems that the configuration of perimetric authentication for WebLogic should rely on headers to extract username and other info (see [3] and [4]).
Thanks a lot,
/ben
[1] https://groups.google.com/group/shibboleth-users/browse_thread/thread/af5e7caf7c9bbc0b/7933e991cfcbe089?pli=1
[2] https://wiki.shibboleth.net/confluence/display/SHIB2/NativeSPSpoofChecking
[3] http://docs.oracle.com/cd/E21764_01/web.1111/e14395/apache.htm#CDEBCGBI
[4] http://docs.oracle.com/cd/E21764_01/web.1111/e13718/ia.htm#DEVSP248 ==> Passing Tokens for Perimeter Authentication

Hi, thanks for the answer.
However, this is not what I meant with "environment variables". I'm not referring to system environment variables, but to session environment variables of the Apache HTTP server, that it can pass along with requests to a Java container. This mechanism is possible with a Tomcat or Jetty container placed behind a Apache front-end, thx to AJP.
I'm wondering how to do this with Weblogic, which does not support AJP.
/ben

Is OHS 10.1.3 compatible with OAS 10.1.2?

I have an existing 10.1.2 standalone OHS and 10.1.2 OAS, and an existing 10.1.3 standalone OHS and 10.1.3 OAS running on different servers. I'd like to build a new server with 10.1.2 OAS and 10.1.3 OAS on it, but with a single installion of standalone OHS in front of them. I currently have OHS standalone 10.1.3 installed, has anyone had luck with this configuration? I'm not sure if my problem is a misconfiguration or if the problem is the topology.
I have 10.1.2 OAS and 10.1.3 OHS clustered, but in the OHS I'm seeing "N/A" for the protocol for the processes, and in the oc4j-status from the web server I'm not getting any processes (if I turn on the 10.1.3 OAS of course it all works as expected).
OAS 10.1.2: opmnctl status @farm -l
-------------------------------------------------------------------------------------------+------
ias-component | process-type | pid | status | uid | memused | uptime | ports
-------------------------------------------------------------------------------------------+------
OC4J | home | 6449 | Alive | 231080160 | 40648 | 22:49:42 | ajp:12501,rmi:12401,jms:12601
OHS 10.1.3: opmnctl status @cluster -l
---------------------------------------------------------------------------------------------------------+------
ias-component | process-type | pid | status | uid | memused | uptime | ports
---------------------------------------------------------------------------------------------------------+------
OC4JGroup:OC4J | OC4J:home | 6449 | Alive | 231080160 | 40648 | 22:49:38 | N/A:12601,N/A:12401,N/A:12501
Note that the protocol is "ajp" when seen from the AS, but the web server is calling it "N/A". When I look up the /oc4j-status page for mod_oc4j it's seeing no processes:
mod_oc4j Global Configuration
hostname : newserver.domain
local instance : IAS-1
select method : Round-Robin
select affinity : None
routing mode : Static
routing ID : g_rt_id
OC4J Dynamic Routing
OC4J Static Routing
destination : instance://oas_1012.newserver.domain:home
context : /something
process :
IASPT Process List
OC4J Process List
process,ias instance,host,port,status
Any ideas? I'd love to hear if anyone has had success with this setup. I'm assuming that OHS 10.1.3 works with OAS 10.1.2, but that could be my big mistake.

Hi Walter,
I never tried your setup..But sound interesting :)
Found few of the oracle doc which will be helpful for you.
http://www.oracle.com/technetwork/middleware/ias/ohs-101310-faq-131717.pdf
From the above link
Can an OHS of Oracle Application Server 10g Release 2 (10.1.2) be used to route requests to Oracle Application Server 10g Release 3 (10.1.3.1.0) instance?
Yes. You can configure an OHS of Oracle Application Server 10g Release 2 (10.1.2) middle-tier instance to communicate with Oracle Application Server 10g Release 3 (10.1.3.1.0) instance.
http://docs.oracle.com/cd/B31017_01/core.1013/b28940/reconfig.htm#BABIEHFI
From the above link
you can use the Oracle HTTP Server from an Oracle Application Server, Release 2 (10.1.2), middle tier as the front-end for your Oracle Application Server, 10g Release 3 (10.1.3.1.0), middle tier.
Please Take a look at below MetaLink note as well.
ID:*260449.1* Everything You Wanted to Know About the Apache-Based OHS Version
Regards
FAbian

How to install webgate on tomcat with solaris 10 AMD 64bits?

Hi Experts
I need to install a webgate on the webserver tomcat, but i found that installation is not possible because thes is not webgate for tomcat, many people recommended install a reverse proxy with Apache or OHS but i can't find a webgate installer on Solaris 10 AMD 64bits.
Exists that version of webgate? or exist a workaround??
what is your suggestion??

Choose the Webserver and the WebGate from the OAM support matix.
If OAM does not have WebGate for the platform where you have tomcat installed, if it is an option, you can install a webserver and WebGate on any OAM supported OS and configure Reverse Proxy. WebGate does not need to be on the same server/machine or on same platform.

OES: How to install WebLogic SSM?

Hi,
I have a small test configuration consisting of:
- IDM (for OID) and IDAM (for OAM) 11g
- a test Apache (not OHS) instance with a 10g webgate, proxying to (using the WebLogic plugin) - this is configured for X509 authentication
- a test WebLogic instance with the OAM ID Asserter and a small test portal app
I have the above configuration working, and I can protect/access control both static resources on the Apache and the portal app on the WebLogic (via the ID Asserter).
I've also stood up an instance of Oracle Entitlement server/OES on a separate WebLogic domain, and that seems to be working, by itself.
Now, I want to integrate these pieces, i.e., I want to be able to control access to the portal app using OES.
I understand that I have to install a "WebLogic SSM" into the security realm on the target WebLogic domain, but I'm having difficulty figuring out how to do that. My understanding is that what's eventually involved is that some additional ATZ and role mapping provider(s) will be added to the security realm, and actually, I'd really like to be able to just drop the appropriate JARs into the mbeantypes directory, then configure the necessary providers in the security realm, e.g., via the WL console, as this would give me a better understanding of what's going on.
Has anyone done something like that? If so, can you provide the "steps" to accomplish that?
Thanks,
Jim

Hi Matt,
Thanks for that info on the options.
I'd really prefer to integrate the WLS SSM into the existing WL domain, but I'm unclear about where you said "That should give you API access to OES within a WebLogic application.".
If I take the 2nd approach you suggested, creating a Java SSM using ConfigTool, etc., are you saying that we'd have to have code in our portal app code specifically to have things work with OES (policies, protections, etc.)? Does it (OES+SSM) not work like other providers in the WebLogic security framework (Authentication providers, Authorization providers, etc.), but rather our app code actually has to be aware of OES and call OES APIs in order to work with OES in-place?
I will also take a look at your 1st suggestion, i.e., running ConfigTool to create a new WL domain, and see what that looks like. In my test environment, I think that will be ok, but in our real environment, where we already have the WL domain(s) established and configured, I think that this (creating a new WL domain, just so that we can integrate OES/SSM) will probably be a no-go longer-term :(, as you can probably understand, since those domains (and security realms) have a lot more stuff in them already than I have in my test environment. Plus, in my test environment, I can blow things away and recover back fairly easily, but we don't have that luxury in production.
Thanks again,
Jim

XDOLoader on RAC / Load balancer

Hi all,
I'm busy with creating an installation script for installing all kinds of objects.
There are also some XML Publisher objects that need to be installed trough XDOLoader.
This installation script has to be dynamic as there are 9 environments that are all 9 in RAC / Load balancing.
Now the problem is that XDOLoader requires an JDBC connection string and I have to 'build up' is this string in my script.
Some will propose to use $TWO_TASK for the SID. And yes, that's a good proposal, but as I'm on load balancing the $TWO_TASK environment variable does not contain an SID...
Also the JDBC_URL from the dbc file does not work.
Does anybody has some experience with using XDOLoader in an RAC / Load balanced environment?
Or any proposals how to get your JDBC string dynamically builded without using $TWO_TASK ?
Thanks in advance
Regards,
Stijn

Hi,
Can you please give me some more information about your setup ...
Something like ...
F5/LB --> Apache or OHS --> WebLogic Cluster..
If you are using F5/OHS directly to WebLogic cluster you can try enabling the session persistence and this problem should go away..
If you are using plain Apache or OHS try putting "DynamicServerList" off.
Let me know something more about setup so that i can contribute to help ..thanks.

Portals, Disco and VPD

If I set up a 10g Portals, Disco environment with SSO, how can VPD be used? I've seen it referred to a few times on this forum, and I'm wondering if it creates some identifier for a database user logging in to Portals or Disco, and therefore a folder or view could limit on this identifier?
Any thoughts on this?
Thx.
Russ

Hi Steve
I was always told that using a VPD with public connections and Portal / Viewer is a very dangerous thing to do if web cache is in use. Of course if we turn webcache off and make sure that the VPD uses the CLIENT_IDENTIFIER then all queries will run every time they are called and be constrained via VPD.
One of the great advantages of having web cache is that it will remember the data from the last time the query was run or from when the portlet was refreshed. But if you are using a VPD, neither Discoverer Viewer nor Portal (Plus does not use web cache anyway) has any idea of what the database is up to and thus the data will be retrieved from the cache - which will certainly break the security because the new user will see the data from the previous execution. Thus, pulling data via a public connection with SSO in use must be restricted to only that data which cannot change.
On page 802 of my Discoverer 10g Handbook, I quote:
It is vitally important that you not attempt to use Web Cache with sensitive data. This is because Web Cache has no mechanism to protect its content. When a request is made to Web Cache for a page that is cacheable (as defined by the caching rules) but has not yet been cached, then Web Cache makes a request on behalf of the user to the Oracle HTTP Server (Apache). OHS has the mod_osso add-on module, whose job it is to guarantee that a user must be authorized to view a page.
The first time such a request is received, the Single Sign-On connection screen will appear and the user will need to provide his or her SSO user name and password. But, because the page is cacheable, Web Cache will keep a copy of it in memory. Then, any subsequent request from any user will deliver the user the page without authentication. This is why Oracle explains that you must cache only Discoverer Viewer content delivered through a Discoverer Public connection. The long and the short of this is that you should not cache pages in Web Cache unless it is okay for everyone to see them.
In other words - do not use a VPD when Web Cache is in use.
I hope this helps
Regards
Michael

FRM-92103: A network error or server failure has occurred.

Hello,
I always get this error after everytime i run a report my forms getting crash.
heres the jre console error :
Java Plug-in 1.6.0_33
Using JRE version 1.6.0_33-b05 Java HotSpot(TM) Client VM
User home directory = C:\Users\charles
Forms Session ID is formsapp.2
The proxy host is null, and the proxy port is 0.
Native HTTP implementation is being used for the connection.
The connection mode is HTTP.
Forms Applet version is 11.1.2.0
FRM-92103: A network error or server failure has occurred.
You will need to restart your application.
Java Exception:
java.io.IOException: Server returned HTTP response code: 503 for URL: http://application:8888/forms/lservlet;jsessionid=yjc6QfjGNkr8fjlyD26fZjqmwBGKkGVgG1L7s91X8y6rBsgGl650!-1255936909
     at sun.net.www.protocol.http.HttpURLConnection.getInputStream(Unknown Source)
     at oracle.forms.net.HTTPNStream.doPost(Unknown Source)
     at oracle.forms.net.HTTPNStream.doFlush(Unknown Source)
     at oracle.forms.net.HTTPNStream.flush(Unknown Source)
     at java.io.DataOutputStream.flush(Unknown Source)
     at oracle.forms.net.StreamMessageWriter.run(Unknown Source)
the http ohs1 server always hang i need to restart it again.
any help?

Michael Ferrante (Oracle) wrote:
I'm sorry, but I'm having a difficult time understand what you are saying...
What do you mean here? "... +if the forms run first then the report will not run it gives always this error+ ..." Are you loading the report results in the same browser that is running the form?
this will be the problem i should posted :
*2.1 Oracle HTTP Server Returns 503 Apache Bridge Failure Error While Toggling Between Forms and Reports*
While toggling between Oracle Forms and Oracle Reports applications through Oracle HTTP Server, you may see the "503 Apache Bridge failure" error. This is due to the wrong patch 12632886 being shipped with the Oracle Forms and Reports installer.
This document found in : http://docs.oracle.com/cd/E24269_01/doc.11120/e25445/install_upgrade.htm#BABHBABB
for those with oraclesupport find the document subject : Forms / Reports 11g Intermittent FRM-92103 Or "Failure of Server APACHE Bridge" - OHS Log Shows "apr_socket_connect call failed" [ID 1380762.1]
I downloaded the patch from
OPatch 12632886 (OHS-WLS COMMUNICATION NOT WORKING ON IPV6 ENABLED SOLARIS M/C [Patch ID 14069292]) for your Windows Platform from My Oracle Support
i successfully patched now its working.
thank you for your patience and help..
Edited by: ck on Aug 7, 2012 3:41 AM

Session Stickness problem on weblogic load balancer

Hello All
I having trouble related to session stickiness for weblogic load balancing, am using httpclusterservlet on a proxy server that forward requests to 3 managed server that are members of the same cluster.
I tried to search for a solution to this problem and I found two links that are talking about the same problem but there are still not answered!!.
Here are the links:
https://forums.oracle.com/message/11158966#11158966
https://forums.oracle.com/message/11158964#11158964
Based on oracle link Using Sessions and Session Persistence, the documentation says under the "HTTP Session Properties" section that there is a change in how the sessions are formatted and this might cause a problem
for some load balancers (the documentation did not say if weblogic httpclusterservlet is one of them), I tried to do the suggested workaround and I got "java.lang.IllegalStateException: Response already committed" error same as what happens with the user in link https://forums.oracle.com/message/11158966#11158966.
Does anyone knows how to fix this??, I search for a very long time on this problem and with no luck, please help!

Hi,
Can you please give me some more information about your setup ...
Something like ...
F5/LB --> Apache or OHS --> WebLogic Cluster..
If you are using F5/OHS directly to WebLogic cluster you can try enabling the session persistence and this problem should go away..
If you are using plain Apache or OHS try putting "DynamicServerList" off.
Let me know something more about setup so that i can contribute to help ..thanks.

Webgate not intercepting when port defined in WLS

We have seen an issue -- thinking it is apache related -- that when the client deploys their application on WLS without using a managed server and directly on the AdminServer... then the webgates do not appear to be catching the app request. This is especially true with SSL (port 443) since 443 is often used as the SSL port for AdminServer.
We have the apache server set up with the following in the httpd.conf file along with the webgate includes:
<IfModule mod_weblogic.c>
WebLogicHost <ip address>
WebLogicPort 7001
</IfModule>
<Location /AppA>
SetHandler weblogic-handler
</Location>
We did not notice this until a client was using 443 as their SSL port for AdminServer. Most clients use the default of 7002 for the AdminServer and then use 443 within their apache or OHS setup. We set the Listen ports to 80 and 443 within the httpd.conf file -- also completed the creation of certs for the webgates to use SSL Ports. If we have the user change their WLS ssl port to 7002, then the webgate catches the request and redirects to the SSO login. Otherwise, the request goes directly to the application with no SSO intervention.
This concerns us since we thought we had the reverse proxy setup correctly -- which means that other client may have the same issue.
Is it possible we are characterizing this incorrectly? Is this a reverse proxy issue or just a configuration issue? It appears that whenever the WLS ports are part of the url (SSL and non-SSL) that the webgate is bypassed. If the non-SSL port is the default 7001 and we include the host/port combination in the HostID and then the URL is http://<hostname>:7001/app... the webgate does not get accessed...
Therefore it is not just SSL ports, but non-SSL as well...

this is a more SSL config relevant.
have you ever tried with the OHS/OAMwebagent to protect the app on WL SSL 7002?
The same story for port 443, I think.
There are more involved in the configurations.

Not able to scale more than 1500 users when using apache (OHS),why ?

Hi,
When scale more than 1500 users, my applications response time degrades.If i use IIS as my webserver I am able to scale more than 3000 users, is there any bottleneck in the oracle apache (OHS) ?
OHS Config
Timeout = 300
KeepAlive = ON
MaxKeepAliveRequests = 100
KeepAliveTimeout = 15
MaxRequestsPerChild = 0
ThreadsPerChild = 5000
SendBufferSize = 16384
MaxClients = 5000
Regards
Vino joy

Hi Jean,
Our application contains customised retail webcenter portal on which retail application is hosted.This portal application is not able to scale up more than 10 users.I had created portal application without the customised framework and then for a similar usecase it is able to scale upto more than 50 users but once it hits 100 VU mark it starts giving errors.If it were an issue with client tokens or login,I would have been able to have more than 1 user at a time.It seems to be an issue with custom framework developed on top of webcenter portal.I have faced a similar issue while testing for this app with jmeter also.I was able to test for the webcenter portal app without any customisation for more than 100 concurrent users. But I was not able to test for more than 1 user for customised retail webcenter portal
Thanks,
Ritesh
Edited by: user766882 on Aug 21, 2012 8:25 AM
Edited by: user766882 on Aug 21, 2012 8:33 AM

SOA suite 10.1.3.3 with OHS Apache 2.0 ... mod_oc4j errors

Hi all,
I have a SOA Suite 10.1.3.3 environment fronted by Apache 2.0.
Installed OHS 2.0 from the Companion CD in a new OHS ORACLE_HOME
Installed SOA Suite 10.1.3.3 in a new J2EE ORACLE_HOME
Can access the consoles ok (BPEL, ESB, etc ...).
Can access EM ok.
Can deploy BPEL processes OK.
However, deployment of ESBs fail.
The Apache 2.0 error_log has the following message:
[Thu Feb 28 10:08:08 2008] [debug] /ade/aime_asrc2_2/oracle/asg_apache/oc4j/src/oc4j_ajp13_worker.c(1288): mod_oc4j: AJP error talking to OC4J atsapdod03:12511: recv failed, 0
[Thu Feb 28 10:08:08 2008] [error] [client 10.38.75.180] mod_oc4j: request to OC4J atsapdod03:12511 failed: Connect failed
Looking at the oc4j-status page, the OC4J container running on atsapdod03:12511 is the container that holds the esb-dt application.
Has anyone out there got Apache 2.0 working with the SOA Suite?

Looks definitely to be an issue with OHS Apache 2.0.
* Shutdown OHS Apache 2.0
* Installed OHS 1.3 in a new ORACLE_HOME.
* Reconfigured OHS 1.3 to run on the same ports as OHS Apache 2.0
* Started OHS 1.3
* All works OK

OHS(Apache 2.0) on RHEL5.3

Hi,
I was trying to install OHS using SOA 10.XX Companion CD got errors like below
httpd/oracle/as/10.1.3/ohs/bin/httpd: error while loading shared libraries: libgdbm.so.2: cannot open shared object file: No such file or directory
httpd/oracle/as/10.1.3/ohs/bin/httpd: error while loading shared libraries: libdb-3.3.so: cannot open shared object file: No such file or directory
Also was OHS(Apache 2.0) on RHEL5.3 is certified?
Thanks,
Bhanu Chander.

Hi,
Try changing the below open file limits set in include/linux/limits.h and include/linux/fs.h (paths based on your OS).
NR_OPEN = maximum number of open files per process
NR_FILE = total number of files that can be open in the system at any time
Hope this helps you.

Upgrading 10g OHS to an apache 2 - mod_so

Hi everyone,
we are running a 10g application server and we wanted to upgrade the apache 1 based OHS to an apache 2 based OHS. The reason is that we wanted to make use of the apache mod_so module but this is not compiled in by default on apache, but it is apparently compiled with apache 2.
Can you give me some guidelines on where i can download the source for OHS and compile it? Or alternatively how i can move from an apache 1 to an apache 2 based OHS ? or what other course of action i can take to have mod_so available?
Thank you very much for your time

Well, just install it, for the steps I think youshould get the installation guide from apache.org, what you will have to do is to configure this apache in a diferent port and at the end, add this url as a Origin Server in Webcache, so it could be accessible.
Sorry but I can't give you much info about how to install it, as I'm not an expert, but I have configure externals apaches to be accesible trew Oracle WebCache, and is at simple as I told, just add it as a origin server.
Greetings.

Trouble starting Apache OHS

New installation of Oracle 11.1 on Dell RHEL/Linux box. Installed Oracle OHS from here into a separate ORACLE_HOME (=db OH/apex).
opmnctl startall starts up opmn but times out on starting Apache OHS httpd. No errors in HTTP_Server~1.log, the last line is simply execing httpd. More surprisingly, no warning/error messages in ohs/logs/error_log, nothing. opmnctl just hangs for a while and comes back with "time out while waiting for a managed process to start". And then I see the opmn process and one httpd process but port 7780 doesn't respond to requests.
Never had any trouble with this before. Everything seems to be setup right but OHS just doesn't want to startup.
Any ideas? Thanks

No, I am not running SSL. My opmn.xml is the default one, no changes.
OK I found the problem. I did a strace -f opmnctl startall and saw that it was waiting on a blocking read on */dev/random*. Some googling on this shows that an insufficient entropy pool (my poolsize had dropped by 62, below the threshold of 64) is a known issue on headless Linux servers due to the lack of truly random activity like mouse and keyboard movement. All this stuff is new to me, it gets really bizarre, thought I was in a physics class. There are some "entropy generating" tools out there but it looked like overkill to me since I am not running a military grade installation. I simply did rm -f /dev/random;ln -s /dev/urandom /dev/random. /dev/urandom generates random numbers of lesser quality but it doesn't block readers. With this fix in place, opmnctl stopall/startall works fine again. Whew, this was a strange problem.
Hope this helps someone else.

Apache or OHS

Similar Messages

Maybe you are looking for