Monday, September 16, 2013

plugin-key.kdb missing

Context

I installed Connections, and it works. I configured the web server, generated and propagated the plugin, and at the time of copying the plug in key store file to web server key store directory, the button is greyed out.





The Websphere HTTP plugin



What are advantages of using webserver and plugin?
The plug-in provides work load management and failover capabilities by distributing requests evenly to multiple application servers and routing requests away from a failed application server.
Static content can be served by the Web server without doing a full round trip to the application server.
The Web server provides an additional hardware layer between the Web browser and the application server, thus strengthening the application server’s security.

Demystifying the IBM® Web ServerPlug-in

plugin-cfg.xml file is created by clicking the “Generate Plug-in” button on the web servers page of the WebSphere administrative console.

HTTP for SSL

 This is all good and useful information. But now more specifically, we go deeper into SSL.

This is a good guide :
Configuring SSL for WebSphere and IBM Http Server : part1

Quote from part 1:
This involves setting SSL for two different communications.
1. Between Browser and IBM http server [IHS]
2. Between IBM http server [IHS] and Websphere Application Server
We are working on 2. The distinction is important.

In my plugin-key.xml :

<Property Name="keyring" Value="c:\IBM\HTTPServer\Plugins\config\webserver1\plugin-key.kdb"/>
<Property Name="stashfile" Value="c:\IBM\HTTPServer\Plugins\config\webserver1\plugin-key.sth"/>

So the keystore should be ready and created, but navigating to the folder it is missing.

But I can see that in websphere, the folder for plugin is :

C:\IBM\HTTPServer\Plugin

And in reality should be :
C:\IBM\Websphere\Plugin
This error is due to the fact that HTTPServer is the default installation folder. So be careful of this when you configure the webserver in Connections 4.5

PLGC0063E

After rectifying (delete and recreating the web server with the good parameters), I can't propagate anymore with errors PLGC0063E and PLGC0049E

PLGC0062I: Le fichier de configuration du plug-in est transmis de C:\IBM\WebSphere\AppServer\profiles\Dmgr01\config\cells\SRVCONNECT1Cell01\nodes\SRVCONNECT1Node01\servers\webserver1\plugin-cfg.xml à c:\IBM\Websphere\Plugins\config\webserver1\plugin-cfg.xml sur l'ordinateur du serveur Web.PLGC0048I: La propagation du fichier de configuration du plug-in est terminée pour le serveur Web. SRVCONNECT1Cell01.SRVCONNECT1Node01.webserver1.PLGC0063E: La transmission du fichier de configuration du plug-in entre C:\IBM\WebSphere\AppServer\profiles\Dmgr01\config\cells\SRVCONNECT1Cell01\nodes\SRVCONNECT1Node01\servers\webserver1\plugin-cfg.xml et c:\IBM\Websphere\Plugins\config\webserver1\plugin-cfg.xml sur le serveur Web a échoué.PLGC0049E: La propagation du fichier de configuration du plug-in a échoué pour le serveur Web. SRVCONNECT1Cell01.SRVCONNECT1Node01.webserver1.

A technote seem to correspond : Plug-in propagation fails with PLGC0063E and PLGC0049E when copying to remote Web server
According to this technote : the message can be ignored if the propagation is successful.

Unfortunately both button are still greyed out.

Copying manually plugin-key.kdb

"Manage keys and certificates" link for plugin-key.kdb is broken in the WebSphere Application Server administrative console

I copied the plugin-key.kdb, plugin-key.sth and plugin-key.rdb from C:\IBM\WebSphere\Plugins\config\webserver1 to C:\IBM\WebSphere\AppServer\profiles\Dmgr01\config\cells\SRVCONNECT1Cell01\nodes\SRVCONNECT1Node01\servers\webserver1

Now the 2 buttons are available.

Friday, September 13, 2013

Missing IBM HTTP Administration Server (Websphere Plugin Configuration Tool)


The IBM HTTP Server is installed but the administration service is missing :



From here I can see that this is related to the configuration in Websphere customization toolbox.

However I can see that in WCT, HTTP Administration Server seems to be enabled :



By redoing the WCT plugin configuration tool, I can find, after enabling putting Administration Server, and where I should I have "run as a service" checkbox :

Non-Administrative user configuration limitation

For those non french reader out there, this basically mean that the user running the configuration is not an administrator, so service cannot be created. Note that this is not mandatory, so that's why I was able to complete configuration without a IBM HTTP Administration service.

My user "icadmin" is in the administrator group and I ran the WCT with administrator rights.

In websphere help I can see that the user launching the configuration should have the rights to "connect as a service"

In Windows "Local Security Strategy" I add "icadmin" to "Act as part of the operating system" and "Log on as a service"


This did not solve.

Similar issues here and here in Connections forums.

I have not been able to solve. I had my rights checked by a Windows admin and he did not find the source of this. This is definitely bizarre.

But following this documentation about Plugin Configuration Tool from Steve Reid and Rob Boretti, I found this command  :
C:\IBM\HTTPServer\bin>httpd.exe -f "C:\IBM\HTTPServer\conf\admin.conf" -k install -n "IBM HTTP Administration for Websphere Application Server V8.0"

The administration service was directly created. It launched ok and I was able to use it from the Websphere Administrative Console.



Monday, September 9, 2013

Connections 4.5 Installation on 2 Windows 2008 Servers Summary


This is a summary for my own use, about installation of Connections 4.5. It's a quick reference documentation, and contains command lines and important points.
This may or may not be understandable from your point of view. I you are beginning at Connections, I suggest that you start with Zero to Hero Guide or Connections Wiki

RAA = Run As Administrator
CLAA = Command Line As Administrator


Server B

  1. For each server, create the administrative user directly on the server (not on the AD domain), add him to Administrators and the group that allow them to use windows at a distance. Log in with this user.
  2. RAA - Install DB2 without a response file. Do not start SSH server automatically. No notifications.
  3. Create LCUSER and add him to DB2USERS group.
  4. CLAA - add licence for DB2, if you have one
  5. CLAA - Configure Unicode
  6. CLAA - Database creation wizard : Run dbwizard.bat then stop and start db2
  7. RAA - Install TDI. Do not specify working directory
  8. CLAA - Install TDI FP7
  9. Modifiy runtime memory of C:\IBM\TDI\V7.1\ibmdisrv.bat
  10. CLAA - Launch population wizard. Just fill the DN of LDAP Bind User. LDAP user search base is empty. "CountryCode" might be set to "null"
  11. Have a cognos administrative account in LDAP
  12. Install DSClient
  13. Install WAS 8 and FP5
  14. Create a typical Application Server profile
  15. Update with ifpm62615 and ifpm71430
  16. Get CognosConfig and copy it on the ServerB. Get db2jcc.jar and db2jcc_license_cu.jar from C:\IBM\SQLLIB\java and put them into C:\Sources\CognosConfig\BI-Customization\JDBC
  17. Modifiy cognos-setup.properties accordingly to your environnement. (Cognos node name : YOURSERVERNode01, installation directories : C:\IBM\CognosBI, C:\IBM\CognosTF and C:\IBM\CognosTF\PowerCubes)
  18. CLAA - Launch cognos-setup.bat
  19. CLAA - Launch cognos-configure.bat
  20. Stop Cognos server and (CLAA) Federate node to ServerA. Then from Dmgr console (serverA) full resynchronize. Start Cognos server.
  21. Validate Cognos Server
Annex : command lines

4. 
db2licm -l
db2licm -a db2ese_o.lic

5.
db2set DB2CODEPAGE=1208
db2start
db2set

8.
c:\IBM\TDI\V7.1\bin\applyUpdates.bat -update TDI-7.1-FP0007.zip
c:\IBM\TDI\V7.1\bin\applyUpdates.bat -queryreg

9.
"%TDI_JAVA_PROGRAM%" -Xms256M -Xmx1024M -Xnojit -classpath "%TDI_HOME_DIR%\IDILoader.jar" %ENV_VARIABLES% com.ibm.di.loader.ServerLauncher %* 

12.
db2 catalog tcpip node db2node remote serverB.domain.com server 50000
db2 catalog database COGNOS as COGNOS at node db2node
db2 catalog database METRICS as METRICS at node db2node

20.
In C:\IBM\WebSphere\AppServer\profiles\AppSrv01\properties\soap.client.props set :
com.ibm.SOAP.requestTimeout=600
C:\IBM\WebSphere\AppServer\profiles\AppSrv01\bin> addnode serverA.domain.com 8879 -includeapps -username wasadmin -password passw0rd


21.
Processes : BIBusTKServerMain, BmtMDProviderMain, cgsLauncher
http://serverB.domain.com:9081/cognos/servlet
http://serverB.domain.com:9081/cognos/servlet/dispatch

Server A


  1. RAA - Install Installation Manager (I use 1.6.2, 32bits)
  2. Install WAS 8 and FP 5
  3. Create a typical Cell profile
  4. Update with ifpm62615 and ifpm71430
  5. At this point I usually have to install Visual C++ runtime library and reboot, else I would get an error at next stage
  6. Install IBM HTTP Server 8.0.0.5 for WAS, Websphere Plugin 8.0.0.5 for WAS and Websphere Customization Toolbox 8.0.0.5 (consists of 3 repository, Webphere Supplements, FP5 for Supplements and FP5 for the Toolbox)
  7. Launch web server plugin configuration tool : 
    • add IHS_Location at C:\IBM\Websphere\Plugins
    • Create a webserver plugin configuration, with ihsadmin for user ID
    • Install IBM HTTP Administration as a service
    • unique server web name : connections and not webserver1
    • local installation
    • AppSrv01 profile
  8. Put shortcuts (Dmgr stop/start, Node stop/start, Administrative console) on the desktop
  9. Start Dmgr and setup Federated Repositories ("Distinguished name of a base entry that uniquely identifies" should be set to "root" with a Domino LDAP). Set as current. 
  10. Enable Application Security. Configure SSO. Log out, restart Dmgr, check groups and users.
  11. Get db2jcc4.jar and db2jcc_license_cu.jar from C:\IBM\SQLLIB\java in ServerB and put them in C:\IBM\SQLLIB\java
  12. Install Connections (cluster name : CL, server name : connections_server). Stop everything, start nodes, full resynchronize the nodes. Stop everything, start everything, Connections last.
  13. Connections is working!

10.
Case is important

13.




Tuesday, August 13, 2013

addNode timeout error

Context : installing Connections, and federating Cognos Node to Dmgr.


The error is :

ADMU0027E: Une erreur s'est produite au cours de la fédération Read timed out ;
           retour à la configuration d'origine.
ADMU0211I: Les détails de l'erreur peuvent être consultés dans le fichier :
           C:\IBM\WebSphere\AppServer\profiles\AppSrv01\logs\addNode.log
ADMU0026I: Une erreur s'est produite au cours de la fédération ; retour à la
           configuration d'origine.
ADMU0113E: Fin du programme avec l'erreur
           com.ibm.websphere.management.exception.AdminException:
           com.ibm.websphere.management.exception.ConnectorException:
           ADMC0009E: Le système n'a pas pu effectuer d'appel RPC SOAP :
           invoke, résultant de [SOAPException: faultCode=SOAP-ENV:Client;
           msg=Read timed out; targetException=java.net.SocketTimeoutException:
           Read timed out]
ADMU1211I: Pour obtenir une trace complète de l'échec, utilisez l'option
           -trace.
ADMU0211I: Les détails de l'erreur peuvent être consultés dans le fichier :
           C:\IBM\WebSphere\AppServer\profiles\AppSrv01\logs\addNode.log

In addNode.log :
[12/08/13 10:04:38:867 CEST] 00000000 ProviderTrack I com.ibm.ffdc.osgi.ProviderTracker AddingService FFDC1007I: Fournisseur FFDC installé : com.ibm.ffdc.util.provider.FfdcOnDirProvider@42f47f5a
[12/08/13 10:04:38:913 CEST] 00000000 Ffdc          I com.ibm.ffdc.util.provider.FfdcOnDirProvider logIncident FFDC1003I: Incident FFDC émis sur C:\IBM\WebSphere\AppServer\profiles\AppSrv01\logs\ffdc\ffdc.1639904841534703718.txt com.ibm.ws.management.connector.soap.SOAPConnectorClient.invokeTemplate 846
[12/08/13 10:04:38:992 CEST] 00000000 AbstractNodeC E   ADMU0040E: Exception sur appel du MBean WebSphere:name=AdminOperations,process=dmgr,platform=proxy,node=SRVCONNECT1CellManager01,version=8.0.0.5,type=AdminOperations,mbeanIdentifier=AdminOperations,cell=SRVCONNECT1Cell01,spec=1.0 mergeConfigDataOnDmgr com.ibm.websphere.management.exception.ConnectorException: ADMC0009E: Le système n'a pas pu effectuer d'appel RPC SOAP : invoke

[12/08/13 10:04:39:195 CEST] 00000000 AdminTool     A   ADMU0026I: Une erreur s''est produite au cours de la fédération ; retour à la configuration d''origine.
[12/08/13 10:04:48:242 CEST] 00000000 AbstractNodeC E   ADMU0040E: Exception sur appel du MBean WebSphere:name=AdminOperations,process=dmgr,platform=proxy,node=SRVCONNECT1CellManager01,version=8.0.0.5,type=AdminOperations,mbeanIdentifier=AdminOperations,cell=SRVCONNECT1Cell01,spec=1.0 doUnMergeConfigDataOnDmgr javax.management.MBeanException: Exception thrown in RequiredModelMBean while trying to invoke operation doUnMergeConfigDataOnDmgr


Here, it is recommanded to increase timeout.

soap.client.props is here : C:\IBM\WebSphere\AppServer\profiles\AppSrv01\properties

Modified timeout to 600.
#------------------------------------------------------------------------------
# SOAP Request Timeout
#
# - timeout (specified in seconds [default 180], 0 implies no timeout)
#
#------------------------------------------------------------------------------
com.ibm.SOAP.requestTimeout=600

In Dmgr, the node has already been created by the addNode.bat command. In order to-- launch the addNode command again, wa have to remove it.

In Dmgr :


However :
  • Removing node from the Dmgr asks the the node agent to be started
  • Starting nodeagent on cognos server returns a fileNotFoundException. Server.xml is not found!
  • It not possible to removeNode using the command line on Cognos server. Error : the node is not federated to a cell.
So it is not usable but not federated. It already exists in Dmgr, but I'm not able to start the nodeagent.

C:\IBM\WebSphere\AppServer\profiles\AppSrv01\bin>startNode.bat
ADMU0116I: Les informations sur les outils sont journalisées dans le fichier
           C:\IBM\WebSphere\AppServer\profiles\AppSrv01\logs\nodeagent\startServ
er.log
ADMU0128I: Démarrage de l'outil avec le profil AppSrv01
ADMU3100I: Lecture de la configuration du serveur : nodeagent
ADMU0111E: Fin du programme avec l'erreur : java.io.FileNotFoundException:
           C:\IBM\WebSphere\AppServer\profiles\AppSrv01\config\cells\SRVCONNECT2
Node01Cell\nodes\SRVCONNECT2Node01\servers\nodeagent\server.xml
           (Le chemin d?accès spécifié est introuvable.)
ADMU1211I: Pour obtenir une trace complète de l'échec, utilisez l'option
           -trace.
ADMU0211I: Les détails de l'erreur peuvent être consultés dans le fichier :
           C:\IBM\WebSphere\AppServer\profiles\AppSrv01\logs\nodeagent\startServ
er.log
C:\IBM\WebSphere\AppServer\profiles\AppSrv01\bin>syncNode.bat srvconnect1.alteca
.fr 8879
ADMU0116I: Les informations sur les outils sont journalisées dans le fichier
           C:\IBM\WebSphere\AppServer\profiles\AppSrv01\logs\syncNode.log
ADMU0128I: Démarrage de l'outil avec le profil AppSrv01
ADMU2026E: Le noeud SRVCONNECT2Node01 n'est pas intégré à une cellule.
C:\IBM\WebSphere\AppServer\profiles\AppSrv01\bin>removeNode.bat
ADMU0116I: Les informations sur les outils sont journalisées dans le fichier
           C:\IBM\WebSphere\AppServer\profiles\AppSrv01\logs\removeNode.log
ADMU0128I: Démarrage de l'outil avec le profil AppSrv01
ADMU2001I: Début du retrait du noeud : SRVCONNECT2Node01.
ADMU2026E: Le noeud SRVCONNECT2Node01 n'est pas intégré à une cellule.
ADMU0211I: Les détails de l'erreur peuvent être consultés dans le fichier :
           C:\IBM\WebSphere\AppServer\profiles\AppSrv01\logs\removeNode.log

But if you simply force delete the Node in the Dmgr, addnode can run. 

Also uninstall the Cognos application, so you start clean. Without this you won't be able to access Cognos.

Success



Friday, August 9, 2013

-classpath is not recognized as an internal or external command operable program or batch file

During Connections 4.5 installation, at the point of launching cognos-setup.bat, I get this error :


Success to verify the JDBC connection to Cognos Content Store database.
'-classpath' n'est pas reconnu en tant que commande interne
ou externe, un programme exécutable ou un fichier de commandes.
Failed to verify the JDBC connection to Metrics database. Please check the error
 message.
Validation failed, unable to continue setup
Which in English is :
-classpath is not recognized as an internal or external command operable program or batch file
 May this be because of parenthesis in my password ?

-> No I changed every related password on this server and no amelioration.

What annoys me is that it's not even an error message ! That's a bug and that bug was already on version 4. I opened a PMR then but finally abandonned.

Finally solved by modifying the cognos-setup.bat!

  1. JDBC Connection for Cognos Content Manager works but not JDBC Connection for Metrics.
  2. I bypass the verifying of the JDBC Connection for Cognos Content Manager and now JDBC Connection for Metrics works ok!
That means that everything is good in my file but the ".bat" is not constructed correctly ??
I have some difficulty believing that this batch was not tested... So what is the difference between my environment and everyone else's ?

Here is the modification of cognos-setup.bat (basically I just added a bunch of "REM") :


REM Validate the JDBC connection.
SET was=%was.install.path%
SET java="%was%\java\bin\java.exe"
SET JAR_HOME=%cognosSetupScriptPath_%BI-Customization\JDBC\*
PUSHD %cognosSetupScriptPath_%lib
SET cp=.;"%CD%\*";"%JAR_HOME%"
POPD
REM Validate the JDBC connection to Cognos Content Store database
REM SETLOCAL ENABLEDELAYEDEXPANSION
REM    set "java_PATH=!java!"
REM FOR /F "tokens=*" %%j IN ('"!java_PATH! -classpath %cp% com.ibm.connections.metrics.cognos.install.CognosDBJDBCConnectionVerifer %cognos.db.type% %cognos.db.host% %cognos.db.name% %cognos.db.user% %cognos.db.password%"') DO (
REM ENDLOCAL
REM IF "%%j" == "JDBC Connection Success" SET cognosjdbcconnection=true
REM CALL :log %%j
REM )
REM IF NOT "%cognosjdbcconnection%" == "true" (
REM CALL :log Failed to verify the JDBC connection to Cognos Content Store database. Please check the error message.
REM EXIT /b 1
REM )
REM CALL :log Success to verify the JDBC connection to Cognos Content Store database.
REM Validate the JDBC connection to Metrics database
SETLOCAL ENABLEDELAYEDEXPANSION
    set "java_PATH=!java!"
FOR /F "tokens=*" %%j IN ('"!java_PATH! -classpath %cp% com.ibm.connections.metrics.cognos.install.MetricsDBJDBCConnectionVerifer %metrics.db.type% %metrics.db.host% %metrics.db.name% %metrics.db.user% %metrics.db.password%"') DO (
ECHO %%j
ENDLOCAL
IF "%%j" == "JDBC Connection Success" SET metricsjdbcconnection=true
CALL :log %%j
)
IF NOT "%metricsjdbcconnection%" == "true" (
CALL :log Failed to verify the JDBC connection to Metrics database. Please check the error message.
EXIT /b 1
)
CALL :log Success to verify the JDBC connection to Metrics database.
CALL :log ... performing validation check completed
ENDLOCAL
GOTO :EOF

During cognos-configure.bat, same problem of course. This time I took a different approached. I duplicated a paragraph which is used to set parameters. This one :

REM Validate the JDBC connection.
SET was=%was.install.path%
SET java="%was%\java\bin\java.exe"
SET JAR_HOME=%cognosSetupScriptPath_%BI-Customization\JDBC\*
PUSHD %cognosSetupScriptPath_%lib
SET cp=.;"%CD%\*";"%JAR_HOME%"
POPD

Here is the final result for cognos-configure.bat :

REM Validate the JDBC connection.
SET was=%was.install.path%
SET java="%was%\java\bin\java.exe"
SET JAR_HOME=%cognosSetupScriptPath_%BI-Customization\JDBC\*
PUSHD %cognosSetupScriptPath_%lib
SET cp=.;"%CD%\*";"%JAR_HOME%"
POPD
REM Validate the JDBC connection to Cognos Content Store database
SETLOCAL ENABLEDELAYEDEXPANSION
    set "java_PATH=!java!"
FOR /F "tokens=*" %%j IN ('"!java_PATH! -classpath %cp% com.ibm.connections.metrics.cognos.install.CognosDBJDBCConnectionVerifer %cognos.db.type% %cognos.db.host% %cognos.db.name% %cognos.db.user% %cognos.db.password%"') DO (
ENDLOCAL
IF "%%j" == "JDBC Connection Success" SET cognosjdbcconnection=true
CALL :log %%j
)
IF NOT "%cognosjdbcconnection%" == "true" (
CALL :log Failed to verify the JDBC connection to Cognos Content Store database. Please check the error message.
EXIT /b 1
)
CALL :log Success to verify the JDBC connection to Cognos Content Store database.
REM Validate the JDBC connection.
SET was=%was.install.path%
SET java="%was%\java\bin\java.exe"
SET JAR_HOME=%cognosSetupScriptPath_%BI-Customization\JDBC\*
PUSHD %cognosSetupScriptPath_%lib
SET cp=.;"%CD%\*";"%JAR_HOME%"
POPD
REM Validate the JDBC connection to Metrics database
SETLOCAL ENABLEDELAYEDEXPANSION
    set "java_PATH=!java!"
FOR /F "tokens=*" %%j IN ('"!java_PATH! -classpath %cp% com.ibm.connections.metrics.cognos.install.MetricsDBJDBCConnectionVerifer %metrics.db.type% %metrics.db.host% %metrics.db.name% %metrics.db.user% %metrics.db.password%"') DO (
ENDLOCAL
IF "%%j" == "JDBC Connection Success" SET metricsjdbcconnection=true
CALL :log %%j
)
IF NOT "%metricsjdbcconnection%" == "true" (
CALL :log Failed to verify the JDBC connection to Metrics database. Please check the error message.
EXIT /b 1
)
CALL :log Success to verify the JDBC connection to Metrics database.
CALL :log ... performing validation check completed
ENDLOCAL
GOTO :EOF

This worked also, but I don't like having to modify the batch. I could have opened a PMR but that would take me 3 weeks again and a lot of work


Wednesday, August 7, 2013

SSO not complete

SSO is only working one way! I login to Connections, then go to webmail, SSO takes care of authentication. But the opposite, when I first login to webmail, it does not log me in to Connections automatically.


Several steps I didn't take, according to Zero To Hero Integration Guide

  1. Editing Wimconfig in order to exernalName instead of uniqueName for userUniqueIdMapping
  2. importing SSL certificate from Websphere

WIM


By the way, what is WIM ? (from What's new in version 6.1? ) :
WebSphere Application Server V6.1 also includes WebSphere Identity Manager (WIM) (also included in WebSphere Portal, which provides basic identity, profile, and user information that can be used by JAAS). Figure 8 illustrates the WIM framework.
So it in linked to authentication.



The wimconfig.xml procedure is explained (because I have to always know why I do what I do!) : here at Kenio blog 

Here is the paragraph I'm referring to :

Please note - if you make subsequent changes to the Global Security Federated Repository area using the ISC - Step 3 may need to be redone as changes may be lost.

What this does -

Step 1.) Insures that the username in the LTPA token created from Domino map to an existing repository in WAS - If there is no match, you get the "user not in defined realm" error in the logs.

Step 2.) Insures that Domino Flat groups can be found for policies

Step 3.) Insures that the username in the LTPA token that WAS generates is resolvable by the Sametime Community Server. In general, Domino does not validate the usernames contained within the LTPA token, it grants the user "default" level access to the database based on the validity of the token.

That seems important! However from that source : http://www-01.ibm.com/support/docview.wss?uid=swg1PM33575 it seems that it is included in 8.0.0.5 having externalName or uniqueName is both fine.

Path of wimconfig.xml : C:\IBM\WebSphere\AppServer\profiles\Dmgr01\config\cells\connectionsCell01\wim\config

However


Wiki : Troubleshooting


In Connections Wiki documentation, they also provide with an action to take on wimconfig file for troubleshooting SSO in the context of using flat group in Domino LDAP.

I applied, and that didn't make any change.


SSO D'bugging


From this Interesting document about SSO in the ICS world (sametime in that case)

D'buggin – It's more an art than science -
● Process of elimination – where is the problem originating?
─ Confirm that basic authentication (username/password) works first
─ Confirm with basic browser based tests before attacking Sametime itself
─ From a browser
– access http://<domino>/names.nsf then go directly to http://<was>/stmeetings
– access http://<was>/stmeetings (login) then go directly to http://<domino>/names.nsf
─ Do both tests – what does this tell you?
– First test – Domino created the LTPA token, Second test WAS created it
– If both tests pass – then continue on to Sametime issues
– If both tests fail – 99% of the time, the LTPA keys are not in synch
– If test(s) fail in one direction only – keys are in synch, but something else is off
– Most common reason for failing from Domino to WAS is “user not in defined realm”

And this point also to "ExternalName" and "UniqueName" modification.

Also (Step 3 = Modifying wimconfig.xml) :
Please note: If you make subsequent changes to the Global Security Federated Repository area using the Integrated Solutions Console (ISC), then Step 3 might need to be redone as changes may be lost.
This is really unfortunate, to have to redo wimconfig modification each time federated repository are changed.

However, I have already ExternalName, so this must be ok.


Getting information


Another source is this paper. What is good about this one, is that, even though this is Websphere portal and not application server, it goes in the direction of getting more information. Debugging by checking parameters is ok at the beginning. But when you checked all the basics, and if you are going in-depth for the first time, my opinion is that you have to know how to get information. I cannot go blind for long.

In fact, this paper is part of a serie on SSO which can be useful if you are interested to have more knowledge on ibm LTPAToken SSO.



Which LTPA

Specifically in my environment, I have :
- several domino in the same domain, SSO ok.
- Connections, SSO works only one way
- Sametime System Console with Proxy Server, SSO ok with domino, and one way with Connections
So basically, Connections side is not OK.
However, is it possible that by adding SSO for Sametime System Console, I erased the LTPA Token for Connections ? How is it possible to check which LTPA Token is in place in domino ?

I reimported the LTPA key from Connections to Domino (restart task HTTP) and Sametime, and now SSO works.

Note: when testing Connections SSO, do not use https://connections.server.com/homepage/login
The "login" at the end puts you at the login page even though SSO will log you if you access https://connections.server.com/homepage

Friday, August 2, 2013

LDAP Bind account and disappearing groups in Sametime

On a production environment, the LDAP Bind account is in LocalDomainAdmins group. That's historical but poses a threat to security. However, when removed, LDAP groups appeared as empty for users in Sametime!

Solving


Testing access to LDAP you have :
- ldapsearch which is delivered with lotus notes
- Ldap Browser from Softerra which i use a lot, and it's free

Clic there to access the free tool

With LDAP testing I found that some groups had the problem and other not. This is LDAP related.


Finally

Some groups had reader field and I had not made aware of them. I guess you never know what's in a production environment until you break your head on it!
I just added the LDAP Bind account to reader access, restarted LDAP task and it rolled.