EP 6.0 and HW Loadbalancer F5

Hi,
we have EP 6.0 SP18 on NW2004 SR1 SP18. We have HW Loadabalancer F5 in front of the Portal. Now we have established that all absolute URL (SAP standard and our customized) in Portal doesn't work.
We connect to Portal with https//proshop.test.s-mxs.net and I see the following URL in Browser window: http://proshop.test.s-mxs.net/.... if the Portal is using absolute URL.
We have started with implementations of iRule on F5 but it doesn't work.
So I have implemented (SAP note 812903) over Visual Administrator Tool the following value in
Dispatcher --> Services --> HTTP --> ProxyMappings
50400=(Host:proshop.test.s-mxs.net,Port:443,Scheme:https,Override:true)
Now it works without errors if the requests are sending over F5 but locally - directly on the Portal Server - nothing is working. I don't get anything.
Has anybody another idea to fix that error.
Regards
Sabine

Similar Messages

OC4J and Hardware Loadbalancer

Hi
I have high available structure in which my two machines are attached with hardware loadbalancer. The scheme followed for load balancing is ROUND ROBIN which means that the load balancer will handover the first request to the first machine and second to the second machine and so on. On each machine we have installed Oracle Application server 10.1.3.0 for production purpose. The operating system on machines is Solaris. I have deployed my application on the both machines but the problem i am facing is that my J2EE application is using Session which cannot being handled in the current scenario. Please let me know what settings and changes should be made on both servers that they start sharing sessions with each other and my problem will be resolved.\
regards,
Raza

Hi,
have a look at this:
http://download.oracle.com/docs/cd/B15904_01/portal.1012/b14037/cg_advnc.htm
Step 7 and may be Step 6
Greetings,
Hagen Malessa

LoadBalancer and BIG IP Setting for Workspace/Planning

Hi,
I recently migrated our Planning applications from System 9.3.1 to System 11.1.2.1 successfully. One of the things we setup in this process was a virtual IP Address to which two server IPs were assigned as well as the 19000 port. This was done to divert reduce the load of logins on one server as well as prepare for a server crash.
The virtual IP address was created successfully and the loadbalancer is able to divert multiple connections to a default algorithm set on the loadbalancer hardware.
However, when i login to the Workspace via my IE. I see that under Navigate->Planning->Application, there is a long string that is displayed instead of "Planning". I was asked to ensure "Persistence" is setup on the loadbalancer and it was confirmed that indeed "cookie persistence" was setup. However, the issue still exists.
I also tried to change a few files residing in the Oracle Folder on the Planning server and the issue still exists. Lastly, i changed the "cookie persistence" to "IP Persistence" with "Source address"persistence as the default setting. The issue still exists.
I have asked Oracle for any documentation that recommends the proper setting and unfortunately there is one and it is for "EPMA" and not for Planning (As i assume the issue is with the "Planning" label.
Anyone else who has come across this issue and what is the best way to resolve this?

Hi,
we had also some difficulties with that buttons and now we solved it with a command bar with different buttons and menus coded in vba.
Was a bit tricky but its was the best way to have a lot of functions on a single workbook.
We used the standard analyzer buttons to do maintain the sequences on a hiden sheet and in vba we connected the buttons/menu entries with vba:
For generatig a command bar in vba but in a module:
Set sb = CommandBars.Add("NAME_OF_COMMANDBAR")
    With sb
     .Visible = True
     .Top = 400
     .Left = 400
    End With
Set TopMenu = CommandBars("NAME_OF_COMMANDBAR")
     With TopMenu
     .Controls.Add(Type:=msoControlButton, before:=1).Caption = "SAVE" '4
     .Controls("Sichern").FaceId = 270
     .Controls("Sichern").OnAction = "Tabelle4.BUTTON_11_Click"
End With
Best Regards,
Thomas

Load Balancing and NIC Teaming

Hi! i have been looking through lots of links and none of them actually can fully answer my queries.
I am to do a writeup on load balancing and NIC Teaming, is there any1 that knows what are the commonly used load balancing and NIC Teaming methods, when to use each method, and the advantages and disadvantages of each method and the configuration for each
method!
Sorry its lots of questions but i have to do a detailed writeup!
Many thanks in advance :D

HI
NIC Teaming - On a single server, you will have mutiple NIC. You can Team the NIC so that both NIC will act togather to provide better bandwidth and High avaliblity.
Example : NIC 1 - 1 GB and NIC -2 1 GB so in Team it can act a 2 GB single NIC, If one fails speed will be reduced but it will have HA
Loadbalancing : Two servers hosting same content:
Example : Microsoft.com can be hosted in two or even more servers and a loadbalancer will be used to split load to each server based of the current load and traffic.
No disadvantages

LOADBALANCER & CLUSTER_ISLAND

I have managed to setup a cluster-island of 2 oc4j instances with a shared web-app and a loadbalancer running between apache and the oc4j instances.
However i don't know how to shutdown the loadbalancer and one of the OC4J instances...i can shutdown on oc4j instance but not the other. The error i get is:
"Error: javax.naming.NamingException:
Lookup error: java.net.ConnectException:
Connection refused; nested exception is:
java.net.ConnectException:
Connection refused"
Does anyone know if i'm doing this correctly or if i'm missing some step ?
Also where can you set the number of max http sessions one instance can serve and maintain in memory ?
And finally does anyone know how this performs in a live production environment...
Thanks, any help would be much appreciated
Cheers
Gurinder

TO shutdown the 2 instances use
java -jar admin.jar ormi://localhost:23791 admin welcome -shutdown
There is no such way to shutdown the loadbalancer itself.
There is a performance hit when you are using loadbalancing. However it is worth mentioning that the bigger the session size the bigger the performance hit for loadbalancing. The session state is multicast everytime there is a setAttribute on the session.
Cheers,
Ashok

Coherence and ADF 11g

Has anyone used coherence for ADF BC 11g ?
What are teh steps to follow and how BC components enables for coherence use ?
Why no sample example on ADF 11g and coherence available ? step by step from install, config and ADF BC using coherence kind.
thx
dd

What other ways to boost ADF BC scalability, performance on high traffic/ Multiple user site?
Actually when you want to use second-level caching it only gives a boost for read-mostly classes.
If you have data that is updated much more often than it is read, do not enable the
second level cache. The price of maintaining the cache during updates can possibly outweigh
the performance benefit of faster reads. Furthermore, the second-level cache can be
dangerous in systems that share the database with other writing applications.
So you must exercise careful jugdment here for each class and collection you want
to enable caching for.
When you have a highly multi threaded environment you can always use a cluster
and a loadbalancer. This is probably the scalability (to increase the capacity of
your application) you are looking for.

Catalyst 3550 stack and etherchannel

I wonder if it is possiable to organize gigabit etherchannel from stack of two catalyst 3550 to server. The problem is to connect server's 2-port NIC to both switches, not just one, and have loadbalancing over 2 links while staying connected in case one of the switches goes down. If not, is it possiable with 3750 switches.

Hi,
This is not possible on 3550's as they do not use true stacking feature on that and moreover both the switches have different configs and they do not get Sync. I think it is possible in case of 3750's, though havnt tried myself.
http://www.cisco.com/univercd/cc/td/doc/product/lan/cat3750/12119ea1/3750scg/swethchl.htm#1033981
regards,
-amit singh

CSS and stickness

Hi all,
          I have a question about WL8.1 clustering and CSS loadbalancing. I hope to get some answers here!
          Environment: WLS8.1 SP5, Solaris, Cisco CSS11500, URL rewriting.
          Environment setup: We have 2 boxes hosting the WL cluster. We don't have a web farm in front. We only have Cisco CSS to perform loadbalancing and stickiness. Our application does not support coockies.
          My question is: Does CSS really care about primary and secondary servers? From what I have understood, if we use a h/w loadbalancer like Cisco CSS, we can configure it do make sticky decisions based on "jsessionid" in the URL (since we use URL rewriting) and the CSS forwards client requests to app servers based on their IP addresses? I guess if we were to use the plug-in, then it looks for primary server id to stick to the server?
          In our our case, we need to have the CSS look for "jsessionid" and not prmary server id. Please correct me if I am wrong.

There is something up I noticed it too - the work around I'm using at the moment is to split classes with the same tag into separate CSS and use @import in the main CSS file.
For instance, suppose i have CSS for body and then body.myClass - Safari ignores "myClass" when specified unless it's saved into a separate CSS file.
I first noticed this behaviour in 10.4.4

PCP in RAC environment is providing lots of ORA-20100: File o0000071.tmp

Hi,
We are on 12.0.6 on 2 node RAC and 2 node application
appl1 --all services including concurrent
appl2 --all services including concurrent with PCP and Hardware loadbalancing
db1 and db2 are two rac nodes. running only db and listener.
but my issue is i have configured PCP, provided
appl1:
====
applprod@erpapp1 $ echo $APPLCSF
/nfs_concurent/ERPPROD_erpapp
applprod@erpapp1 $ echo $APPLPTMP
/nfs_concurent/tmp
applprod@erpapp1 $ echo $APPLTMP
/oracle/applprod/inst/apps/ERPPROD_erpapp1/appltmp
applprod@erpapp1 $
appl2:
====
applprod@erpapp2 $ echo $APPLCSF
/nfs_concurent/ERPPROD_erpapp
applprod@erpapp2 $ echo $APPLPTMP
/nfs_concurent/tmp
applprod@erpapp2 $ echo $APPLTMP
/oracle/applprod/inst/apps/ERPPROD_erpapp2/appltmp
applprod@erpapp2 $
=====
and it is mounted onall the Nodes including dbnode.
we have updated the utlfile path of both the db instances with /nfs_concurent/tmp as first parameter in utl_fil
===========================
SQL> show parameter utl
NAME TYPE VALUE
create_stored_outlines string
utl_file_dir string /nfs_concurent/tmp, /usr/tmp,
/nfs_data/ftp/idmin, /nfs_data
/ftp/idmout, /nfs_data/ftp/scb
out, /nfs_data/ftp/scbin, /nfs
data/ftp/citiin, /nfsdata/ft
p/citiout, /nfs_data/ftp/amsou
t, /nfs_data/ftp/amsin, /u02/E
RPPROD, /u01/app/oracle/produc
t/10/db/appsutil/outbound/ERPP
ROD_erpdb1
SQL>
==================================
The issue is both the db nodes are not creating the *.tmp files in sequence they are
RAC node1:
SQL> select applsys.fnd_temp_files_s.nextval from dual;
NEXTVAL
38562
RAC node 2:
SQL> select applsys.fnd_temp_files_s.nextval from dual;
NEXTVAL
39308
select ORDER_FLAG from dba_sequences where sequence_name='FND_TEMP_FILES_S';
ORDER_FLAG
N
If a node 2 is generating the tmp file *309*.tmp which might already created by the other node then the above error is coming to us,
If you shutdown any one node and cleared all the *tmp files from the /nfs_concurrent/tmp then we will not get the error.
In our issue it couldhave been overwrite the earlier file but the issue is we are not having boththe database user ID is defferent, We are in the process of changing the same but we wanted to change our setup, it should create the new no which was not generated by the other node.
Pl help us in achivingthe same..
oracle@erpdb2 $ id
uid=207(oracle) gid=300(oinstall) groups=301(dba),206(citygrp),207(sbigrp)
oracle@erpdb2 $
oracle@erpdb1 $ id
uid=206(oracle) gid=300(oinstall) groups=301(dba),206(sbigrp),207(citygrp)
oracle@erpdb1 $
We have gone through the note "261693.1Troubleshooting ORA-20100 on Concurrent Processing" to trouble shoot the same,
Remind: application is working fine if you run on the single database node..
Regards
Vasu

Hi Husain,
We have gone through the suggested note, in our case the sequence is not jumping, but in both the nodes the sequence is not the same value..
NodeDB1
====
SQL> Select fnd_concurrent_requests_s.nextval from sys.dual ;
NEXTVAL
631070
SQL> /
NEXTVAL
631072
SQL> select FND_S_FILE_TEMP_ID.nextval from sys.dual;
NEXTVAL
8836
SQL> /
NEXTVAL
8837
SQL> select FND_TEMP_FILES_S.nextval from sys.dual;
NEXTVAL
40151
SQL> /
NEXTVAL
40152
SQL>
NodeDB2
===========
SQL> Select fnd_concurrent_requests_s.nextval from sys.dual ;
NEXTVAL
631069
SQL> /
NEXTVAL
631071
SQL> select FND_S_FILE_TEMP_ID.nextval from sys.dual;
NEXTVAL
8844
SQL> /
NEXTVAL
8845
SQL> select FND_TEMP_FILES_S.nextval from sys.dual;
NEXTVAL
40309
SQL> /
NEXTVAL
40310
SQL>
Pl suggest..
Regards
Vasu

CSS11501 VIP just stops after a day or two

We have a CSS11501 that we are using to frontend an oracle colaberation suite cluster of unix servers. There are two servers in the first cluster. I have set it up to have a single load balanced vip with the following content services (http, imap, smtp, port 7777, and port 9401). Two other servers have a different load balanced VIP for their http and LDAP funtions. We are trunking in from a Cat3500XL switch the two vlans (10 "public" and 250 "private server") via the gig interface. It works fine for about a day or two then the VIP just stops responding to any traffic. We have done dumps on the servers and they are not getting any traffic through at all. Each server has a seperate VIP & source group set up as a straight pass through for maint access, these continue to work fine. I have tried stopping and starting the services, content rule, and source groups. even totally deleting the VIP and recreating it does not work. Rebooting the switch was of no help as well. If I recreate all of the content rules on a new VIP and assign the services it comes up fine but the old VIP address just stays dead. It is almost like the VIP has a limited lifespan and then shuts down. Any suggestions would be helpfull as this is a new technolagy for me. Thanks.
Burke McCrory
Internet Administrator
Oklahoma Tax Commission

you will have to capture a sniff in front of the css to see if it is receiving traffic for the vip.
You can also do a 'sho summary' and see if the hit counter is increasing.
Then check the upstream device and verify the mac address associated with the vip.
Make sure the L2 cam table is pointing to the CSS.
Do you have redundancy turned on ?
Finally, not sure what is the colaboration suite exactly, but if this is a software loadbalancing tool, you may want to stop using it if you have a CSS. Simply configure the server independent from each other.
In the past we have seen many issues when trying to combine both software and hardware loadbalancing.
Gilles.

SSL: Best Practice in HA environment

Hi all,
The customer want to know, if there is any best practice Doc available regarding SSL Konfiguration in one WebLogic Server (10.3.6) - HA environment and Hardware Loadbalancer? Additional questions:
1- Configuration of lacation of "certification files"? which practice is better: share Storage or each Managed Server?
2- More Information regarding generating of certificates
Thanks in advance, Moh

Hi Moh,
Did you have a look at this?
http://www.oracle.com/technetwork/database/availability/maa-fmwsharedstoragebestpractices-402094.pdf
Search for certs.
Regards Peter

Analytics server 1.1

I installed analytics server 1.1. Installation went successfully. When I loaded analytics console community, I get javascript error PTHTTPGetRequest is undefined. Any idea what is creating this problem:
I have win2003 server/portal 5.03 .NET/Sql 2000 SP3a.
thanks

I was facing exactly the same problem since Friday I installed Analytics Server on a pre-production platform.As suggested by Eric, I searched into the log files of Analytics to get something wrong.I found
10 Mar 2006 16:08:27,408 FATAL JSIncludesHandler - com.plumtree.openfoundation.io.XPFileNotFoundException: jscomponent file for jsutil not found or failed to load.Exceptions encountered: - com.plumtree.openfoundation.io.XPFileNotFoundException: Failed to load http://img-rec.portal.groupe-casino.fr/imageserver/plumtree/common/private/js/jsutil/jscomponent.xml. Loader exception: Connection refused: connect
In the analyticsui.log file.
My first understanding was that this javascript file was missing or corrupted. So Analytics Remote Server was encountering exactly the same kind of error as the client browser.But I checked that the file was present on the imageserver.
So I tried the exact HTTP request (http://img-rec.portal.groupe-casino.fr/imageserver/plumtree/common/private/js/jsutil/jscomponent.xml) being logged on the Analytics Remote Server. The request did not finished. In my case, I had a problem to access the imageserver (problem with the VLAN configuration and hardware loadbalancer for imageserver).Solving this, there's no more problem.I encourage you to check the accessibility of the imageserver from the Analytics Remote Server.
I was encouraged to search this way because I had one time the same kind of error with the content server. Content Server was requesting for a javascript file of the imageserver and never got it. This resulting with a blocking state for Content Server. In this case, the javascript file was really missing. But I discovered this time that such remote servers as Content Server also request for the imageserver component.
My understanding in the case of the problem with the PTHTTPGETRequest missing function is the following.The PTHTTPGETRequest function is hosted in the PTXML.js javascript file.This javascript file is located in the E:\sysapp\plumtree\ptimages\imageserver\plumtree\common\private\js\jsutil\XXXXXX directory. The beginning of the directory depends on your installation. XXXXXX refers to the current version (something like 157085)This is the same for other fundamental javascript functions for the client.So the client needs to make the appropriate includes of javascript files according to your installation and the products you're using.The name of the files to be included are determined by the Remote Server (Analytics in our case). To get the right version to include, Analytics has to access the http://img-rec.portal.groupe-casino.fr/imageserver/plumtree/common/private/js/jsutil/jscomponent.xml. This jscomponent.xml file contains the id of the current version (157085 for instance). Because the Analytics Remote Server don't get this file, I guess it doesn't know what is the right version. So it doesn't put any include clause into the HTML answer to the client.But for any reason, the Analytics Server put in the HTML answer the code with the PTHTTPGETRequest.In the HTML page, there's the PTHTTPGETRequest call but no include. This leads to the blocking problem.
I also established this analysis thanks to the name of the component logging the error message: JSIncludesHandler. I understood this component is in charge of defining the right javascript includes for the client.
Let me know if this is not clear __

Form on coldfusion page losing form variables

OK, I am having a very strange problem. Several applications
on our
production web servers have started manifesting the same
symptom. On
the login page, you have a couple of fields for account name
and
password. When you hit submit the action page says it can't
find the
form variables. These are on a cluster. If you go to the
servers
individually, then everything works as it should. Our network
people
claim that we are not shifting from one server to the other.
Has anyone
else ever seen this behavior before on a single
server?

These are stand-alone servers that being load-balanced. They
are supposed to stuck on the same server until the session ends. If
we go to the individual servers and run the applications, it works.
Otherwise, it fails randomly.
quote:
Originally posted by:
fober1
Hi,
I have not seen that specific behavior on clustered or single
web-servers.
You did not write if you are using any cluster features of
ColdFusion, but I assume you are using as stand-alone configured
ColdFusion servers, and through LoadBalancer configuration (sticky
cookie) you are making sure that users are always sent back to the
same server.
I would not expect the problem you have to be caused by such
a load balancer configuration, but companies who implement load
balancers often also use Proxy-servers, and now well they can mess
up a lot of things with caching.
(even Dan's problem could be caused by a proxy server: first
request to verify if the requested page still has the same date,
and then the request is actually sent to the server itself for
execution).
A quick way to verify if a proxy is messing with you is to
attach any URL parameter: "www.mysite.com/index.cfm" may be cached,
but "www.mysite.com/index.cfm?user=12345" should not be served from
the cache and instead get sent to an actual web-server to return a
customized response.
Cheers,
fober

Issues jumpstarting solaris 10

Here is a quick overview of my setup:
I have a B1600 shelf with 5 B100n's and 1 LoadBalancing Blade.
I also have a 280r on the same network which is the install/tftp/config server for jumpstart.
I am having a problem trying to jumpstart a new blade with solaris 10. When I "boot net - install" it just does this: Requesting Ethernet address for: (some random ip on the class b network). I do see it accessing the tftp server and pulling the boot image and then mounting the install path to the Solaris 10 cd's. I never see it trying to mount the directory where the sysidcfg file is located though. The solaris 10 jumpstart does work if I do "boot net - text" for a interactive install but I need to do the automated since I am doing a flash install. I guess the problem is that it does not try to pull the sysidcfg file which I don't know why. It is specified in the /etc/bootparams file correctly. Here is my add_install_client line:
/export/install/sol_10/Solaris_10/Tools/add_install_client -e <ether> -c server:/jumpstart -p server:/jumpstart <hostname> sun4u
Using the same syntax, I can jumpstart with solaris 9. Anyone with some help?

Try this link which will give a idea for installng the NIC
http://solaris-x86.org/documents/tutorials/rtls.mhtml

BPF Launch from second application server

I have the following setup.
SAP BPC 7.0 SP9 Patch 1
SQL Server 2005 Enterprise Edition SP2
Two application servers and one database server.
We have a loadbalancer F5.
When I launch the BPF and the loadbalancer takes me to the second application server. Then the message comes up that the templates does not exist. When I go to eTools and try to open the Company folder, first it says that the folder thas not exist and after about 30 seconds the templates start appearing.
I have followed the instructions on how to do the Loadbalancer, and I can see that the balancer is working, but templates take long to appear from the second application server.
It seems like there is a delay in the templates showing.. Do you have any idease why this happens? Or what settings I can have a look at?
Regards
Henry

So first of all I will check the tables:
tblserverinfo and tblappsetinfo.
Into tblserverinfo you must have just records with suffix 1 and for application server and web server you must have virtual Ip of your NLB.
Iinto tblapsetinfo all the records refering servers must you just name with suffix 1. (sql server1, olap server 1...)
The second check I will do it regarding where do you have the file server.
Because you have two application servers then you must have a share drive where you have the file server because the templates must be access from both servers.
So for the file server you will not be able to use local drive because this can be a problem when you try to open the templates.
That's means you have to check if you can access the path specified for file server (you can find it into tblserverinfo) from both application server. It can be that actually from second application server you have access problems.
The third thing to be check is the affinity(persistency) of you NLB F5.
In your case because the afinity must be single (persistency on) because once which you connect to one application server you would like to have all request going to same application server.
All the tests must be done with antivirus and firewall disable into both application servers.
Regards
Sorin radulescu

EP 6.0 and HW Loadbalancer F5

Similar Messages

Maybe you are looking for