Connections

Quick Tip – Name your console..

May 28, 2015Connections, Sametime, WebSphereConnections, Webspherersullivan77 Leave a comment

The guys I work with are full of good tips.

I’d always wondered how to do this, as I thought it was some sort of Developer jiggery-pokery.

Most environments I work in have a Development environment, and a Production environment.

Quite often, I get lost. Am I in Prod or Dev? Especially with cryptic server names.

Quick way to fix this is to Name your Consoles.

Logon to your ISC.
Expand System Administration > Task Management
Select Console Identity.

Type in the Console name..ie Prod

And hey presto…You know where you are..

You’re welcome.

CCM CR2 Upgrade/Backup Script errors – Access is denied.

April 21, 2015ConnectionsCCM, Connections, CR2, IBM Connectionsrsullivan77 Leave a comment

Previous post was how to install CCM. This post will be about an error that I experienced just trying to apply the Fix Pack to get it to CR2 level.

When applying the CR2 update, the Content Engine Server is the first cab off the rank.

Following the instructions, I started with the D:\IBM\Connections\ccm\ccm\ccm\scripts\backup.bat

This script failed to run. Error was ADMN0022E: Access is denied……..because of insufficient or empty credentials.

*sigh*

I checked that the password wasn’t locked out. It wasn’t.

I logged into wsadmin manually, jut to verify all was good. I could.

So google put me onto this awesome post. In this case, wasn’t my issue, although I did try the fix.

Time to dig deeper.

I grabbed the backup.bat file, and scanned though till I found the part that was failing.

I enabled output to the screen by adding echo on..

I then ran the backup script again. This revealed what the issue was.

The customers WAS admin password contains an exclamation point “!” for password quality. This script was passing this password as a variable to the the script which was running the wsadmin command. There appears to be a second time that this script calls the %was.admin.passwd% variable, but as it does it strips the “!” in my password. So I’m supplying the incorrect password, even though it’s correct at the time of input.

I’ve not been able to find a resolution to this apart from changing the password.

I had tried escaping the character out, and used different variations but to no avail. If anyone knows how to do this, please let me know.

CCM Installation fails with CLFRP0038E – Connections 5 CR2

April 17, 2015ConnectionsCCM, Connections, CR2rsullivan77 One comment

I ran into this issue at a customer site I’m working on at the moment.

Here’s the scenario. Installing the following: Connections, Survey’s Polls (FEB), Social Mail, External Access, CCM, File Viewer. I’ve installed Connections, updated to CR2 and verified all is working. I’ve also installed FEB.

Following this documentation, I ensure I have all of the CR2 required files for install, including the fixpacks extracted to a directory. I’m installing CCM across 2 nodes in a cluster, on separate hosts.

I wanted to make sure that the Clusters/Servers were setup correctly in WAS. I have not had much luck with installers creating WAS clusters or servers, so as this is pretty straight forward I like to set them up for installations to use.

Below is the summary screen of the CCM_Cluster.

Time to install.

I’ve checked the prereqs. Note- Make sure you have at least 6GB free in %Temp%. In my case this was located on the C:\

I select the Modify option.

Select Add-on features IBM Connections Content Manager

You’ll be asked for your existing WebSphere credentials. Put them in in then select Validate.

You’ll then hit this screen

I selected New Deployment

Left remaining blank, then selected the Directory where I extracted all the installers and FP’s

If you’ve extracted everything out, you should get a Validation Successful. If not, check what files you are missing. The installer was pretty good in identifying what packages I was missing.

I then selected where I wanted to deploy CCM. I select my Servers/Cluster that I previously setup.

Once this was selected, got to the next screen which I selected Modify.

CCM then installs.

I got the below error.

Damn…

Looking through log files got me this..

Further searching through the logs at <WAS INSTALL>\profiles\Dmgr01\ConfigEngine\log\ConfigTrace.log reveals that a properties file for the nodes/cluster I’d setup did not exist in the D:\IBM\Connections directory.

[echo] Communities server: Cluster1_server1
 [echo] Loading properties from file: D:/IBM/Connections/Cluster1_server1.properties
 [echo] Communities server HTTP Port: 9082
 [echo] CE server: CCM_server1
 [echo] Loading properties from file: D:/IBM/Connections/CCM_server1.properties
--- Exception Thrown ---
D:\IBM\Connections\ccm\ccm\ccm\config\includes\ccm_cust_cfg.xml:1200: Source file does not exist.
at org.apache.tools.ant.taskdefs.LoadProperties.execute(LoadProperties.java:159)

This is where it got weird. I’d thought that is was because I’d setup the nodes/cluster outside of the install and that it was the IBM Installation Manager that created these properties files. Nope. I tried by setting up the nodes/clusters using the installer, same result, CCM not installed, same error.

Fix? Workaround?

To fix this one I created the files manually. This required grabbing the following values from my ISC in WAS.

I created 2 files, one for each of my servers.

These files were created before I hit Modify on the final installation screen, i.e after you’ve done all the above. I’m pretty sure that you could do this before you ran the installer.

Ran the installer again and was able to install.

Hope this helps someone.

wsadmin ProfileAdmin error “NameError: bAskForNodeComm”

March 4, 2015ConnectionsConnections, connections 4.5, fix, howtorsullivan77 Leave a comment

I stumbled across this issue when trying to update a users information in Connections after a UID change.

I’d successfully updated the PROF_UID and PROF_UID_LOWER with the new information, and TDI had synched across the new user data.

To then get the other Connections Applications to update their user information for this user, I ran the ProfilesService.publishUserData(“new.email@domain.com”)

I received a very descriptive error.

So I noticed that the unlike the other Connections admin commands, the ProfilesAdmin.py script had not asked me for “What node did you want to connect to?”

I tried various things, but was not able to get the ProfilesAdmin script to select a node.

I found this error which seemed to match what I was getting, so I tried the suggested fix.

See below, it worked. This technique also will work with other ProfileAdmin commands.

Error when reparenting a sub-community.

February 16, 2015Connections, WebSphereConnections, connections 4.5, howtorsullivan77 Leave a comment

Reparenting a sub-community is a feature added in Connections 4.5 CR3.

This feature allows you to move a sub-community to a Parent level, but also allows you to move the sub community to a different parent. (Think of the sub community as the child).

To move a sub-community, you use wsadmin, and the CommunitiesService.moveSubcommunityToCommunity(“CommunityUNID”) command.

Here is the IBM Link to the this utility.

I ran into an issue when trying to move the Community.

Firing up wsadmin as our usual Admin user, I got the below error.

The System.Out

000009a TangoServiceI W com.ibm.tango.internal.service.TangoServiceImpl getMemberProfileWithUpdates CLFRM0110W: Undetermined memberProfile, in which its name: waslocaladmin, email: null, member uuid: 2f333cc26-b4d5-437d-b970-c9c1b3c076aa, and logins: [waslocaladmin], closely matches to directory service object of an user, whose name: waslocaladmin, email: null, and logins: [waslocaladmin].

Then

000009a TangoServiceI E com.ibm.tango.internal.service.TangoServiceImpl updateCommunity CLFRM0039E: internal error
com.ibm.tango.exception.MemberDuplicateLoginIdException: [waslocaladmin]
at com.ibm.tango.internal.service.TangoServiceImpl.getMemberProfileWithUpdates(TangoServiceImpl.java:3187)

This was interesting, as the user waslocaladmin user was the Admin user for Connections, but wasn’t listed as an Admin userroles in the Application.

Additionally, waslocaladmin user was in the file based repository.

Workaround..

Switching to our support login, which has Administration access to the Communities application, I was able to run the reparenting command successfully.

This was done by running the wsadmin command as the support user.

Fix..

Further Troubleshooting revealed that this issue is more than likely the waslocaladmin id not being synced with the Community member database table.

This environment was upgraded, so in theory this could be correct. The migration method was a side-by-side install, so the waslocaladmin user would be different.

Synchronise a single member’s directory ID in the Communities member database table

Open a command prompt and navigate to C:\IBM\WebSphere\AppServer\profiles\Dmgr01\bin
Run wsadmin -lang jython -user waslocaladmin -password password -port 8879

From the wsadmin prompt run:
wsadmin>execfile(“communitiesAdmin.py”) (enter)
wsadmin>CommunitiesMemberService.syncMemberExtIdByLogin(“waslocaladmin”)

Errors have stopped, so fingers crossed!

Connections theme reverts to 4.5 after applying C5 CR1

January 18, 2015ConnectionsConnections, Webspherersullivan77 Leave a comment

Just a quick post. I ran into this issue, and it’s not the first time I’ve seen it.

After applying the CR1 update to Connections 5, the theme reverted to version 4.5.

I made sure that all nodes were synced, restarted environment and the same issue was experienced.

The Fix

I removed the temporary caching directory for the WebSphere Server running Connections.

This forces WAS to update it’s cache on next restart.

Shut down the Application server running Connections.
Go to the <install_root>/profiles/<servername> directory
Rename the /temp directory to /old_temp (as below.)
Rename the /wstemp directory to /old_wstemp (as below)
Restart Application Server.

Once testing is completed, you can remove the old temp directories.

Moving Databases from DB2 9.7 to 10.1 on Windows 2012 R2 for C3-C5 upgrade.

November 3, 2014Connections, DB2Connections, db2, upgrade, windows_2012rsullivan77 Leave a comment

Those people familiar with Connections will be aware of the different methods to upgrade. The one that we use most often, due to the fact that we have the best chance of success, is the Side-by-Side migration. With this migration you essentially setup a new environment at the version you are upgrading to , then transition your data. This also allows for the old version to be available to the customer, so that historical data is still available.

We mainly use DB2 for the backend, and most of time it’s hosted on a Linux machine. But some times we come across Windows.

This is one of those cases….

I ran into an issue with migration of DB2 servers between versions of DB2, hosts and Operating Systems.
Moving from DB2 9.7 on Windows 2003 – DB2 10.1 on Windows 2012 Server R2.

Firstly, here is an overview of the process used.

On source host – this requires an outage of your Connections host.
Login as DB2ADMIN to the local machine.
Databases are backed up offline using the DB2 Control Center.
Backups are to Files.
On Target DB2 server
Logged as Domain user
Copy from the Source host to the Target host the DB Backups.
Switch to DB2Admin
Restore Databases. Nice thing with this is that the databases were automagically upgraded to 10.1.

The issue

When the databases were restored to the DB2 10.1 server, DB2Admin on the Target machine did not get the correct permissions and authorities to the databases. Even though DB2Admin was explicitly listed in the Database, this did not help.
The problem propagated itself further in the DB2 Upgrade wizards in Connections. When the DBUpgrade wizard ran, it could not verify which Connections version the databases were. This leads me onto the next thing..how do the Connections wizards identify what Connections version the database is?
Each database has a specific table that holds what version of Connections the database is. Being IBM, having these be the same table would be way too easy. The Table for HOMEPAGE is HOMEPAGE.SCHEMA.

When viewing the data in HOMEPAGE.SCHEMA, as DB2Admin, nothing was returned, We could see no data.
DB2Admin had no rights to the database, which makes no sense.. DB2Admin should be god.

The fix

Was to perform the following on the Target DB2 Server before the restore.

Set the DB2_RESTORE_GRANT_ADMIN_AUTHORITIES registry variable BEFORE preforming the restore into a new database.
Example:

db2stop
db2set DB2_RESTORE_GRANT_ADMIN_AUTHORITIES=ON
db2start

Just to automate the restore process, I created a batch script.

 D:
 db2 RESTORE DATABASE BLOGS USER db2admin USING password FROM "D:\DB2TFR2" TAKEN AT 20141022174241 TO "D:" INTO BLOGS WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1
 db2 RESTORE DATABASE DOGEAR USER db2admin USING password FROM "D:\DB2TFR2" TAKEN AT 20141022174313 TO "D:" INTO DOGEAR WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1
 db2 RESTORE DATABASE FILES USER db2admin USING 1password FROM "D:\DB2TFR2" TAKEN AT 20141022174401 TO "D:" INTO FILES WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1
 db2 RESTORE DATABASE FORUM USER db2admin USING password FROM "D:\DB2TFR2" TAKEN AT 20141022180225 TO "D:" INTO FORUM WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1
 REM db2 RESTORE DATABASE HOMEPAGE USER db2admin USING password FROM "D:\DB2TFR2" TAKEN AT 20141022180322 TO "D:" INTO HOMEPAGE WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1
 db2 RESTORE DATABASE OPNACT USER db2admin USING password FROM "D:\DB2TFR2" TAKEN AT 20141022180348 TO "D:" INTO OPNACT WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1
 db2 RESTORE DATABASE PEOPLEDB USER db2admin USING password FROM "D:\DB2TFR2" TAKEN AT 20141022180432 TO "D:" INTO PEOPLEDB WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1
 db2 RESTORE DATABASE SNCOMM USER db2admin USING password FROM "D:\DB2TFR2" TAKEN AT 20141022180459 TO "D:" INTO SNCOMM WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1
 db2 RESTORE DATABASE WIKIS USER db2admin USING password FROM "D:\DB2TFR2" TAKEN AT 20141022180525 TO "D:" INTO WIKIS WITH 2 BUFFERS BUFFER 1024 PARALLELISM 1

I performed the above, restored the Databases. This time the DB’s took considerable time to restore….more stuff happening! I was now able to successfully access the Databases, and run the wizards.
Happy days,

SQL Locking on SNCOMM DB with MS SQL Server

September 28, 2014Connections, SQL ServerConnections, PMR, SQLrsullivan77 Leave a comment

Recently I was asked by a customer to investigate the poor performance of their Connections 4.5 CR3 environment. They had noticed that the servers were slowing, and eventually Connections ground to a halt.

Checking the System.out logs on the InfraCluster servers revealed that issues were being experienced with process hangs and process queuing.We then traced this back to the Databases, as it appeared like anything that relied on Communities were slowly failing. We found our problem.

The SNCOMM database had a lock from a long running task. Analysing the lock showed that this was the select statement it was running:

@P0 nvarchar(4000))   SELECT       TC.COMMUNITY_UUID as COMMUNITY_UUID,   TC.NAME as NAME,   TC.CREATED as CREATED,   TC.CREATED_BY as CREATED_BY,   TC.LASTMOD as LASTMOD,   TC.LASTMOD_BY as LASTMOD_BY,   TC.LOWER_NAME as LOWER_NAME,   TC.PLAIN_DESCR as PLAIN_DESCR,        '' as IMAGE,        '' as DESCRIPTION,   '' as DESCRIPTIONEX,       TC.MEMBER_COUNT as MEMBER_COUNT,      TC.PEOPLE_COUNT as PEOPLE_COUNT,       TC.GROUP_COUNT as GROUP_COUNT,   TC.COMMUNITY_THEME as COMMUNITY_THEME,       '' as CREATEDBY_DIRECTORY_UUID,   '' as CREATEDBY_DISPLAY,   '' as CREATEDBY_EMAIL,    '' as LASTMODBY_DIRECTORY_UUID,   '' as LASTMODBY_DISPLAY,   '' as LASTMODBY_EMAIL,     TC.COMMUNITY_TYPE as COMMUNITY_TYPE,    TC.HANDLE as HANDLE,     TC.PARENT_UUID as PARENT_UUID,   TC.ORG_ID AS ORG_ID,   TC.INTERNAL_ONLY AS INTERNAL_ONLY,   TC.OWNER AS OWNER,   TC.MEMBER_UPDATED as MEMBER_UPDATED,   TC.REF_UPDATED as REF_UPDATED,   TC.IMAGE_UPDATED as IMAGE_UPDATED,       TC.PRE_MODERATION as PRE_MODERATION,       TC.POST_MODERATION as POST_MODERATION   FROM   SNCOMM.COMMUNITY TC       WHERE      TC.COMMUNITY_UUID=@P0

Removing the lock resolved the issue, everything in the environment “trued” and all was well with the world again.

The customer then did some investigation into what changes were made in the environment. Turns out through some pretty awesome sleuth work that a new user had been brought on that day. This user was added to a Community that had >10 sub-communities. This was the clue..

In Connections 4.5 (unsure which version) a new feature to update the membership of sub-communities when a user was added to the parent community was introduced. Running this code for under 10 sub-communities, no problem. More than 10; LOCK.

PMR was logged and the resulting investigation showed that this was a reproducible error. Bad news is that it also exists in version 5, but this will be resolved in an upcoming fixpack.

dontforgetthe0

My journey installing, configuring and supporting IBM Collaboration Software.