Thursday, July 16, 2015

Active-Active Disaster Recovery

In my organization, we are very big on DR site being available . In the event of any outage in the datacenter, the failover and failback processes are required to be nearly seamless. From the end user perspective, It is important for them to launch the same link for Smartview and be able to connect to Essbase, Reports, Planning.

In order to achieve this, we need the following:

Automatic DNS redirection

Any redirection tool of your choice can be implemented in your organization. In my case, both my production and DR VIPs are pointed to a separate dns url which gets the heartbeat information and redirects traffic to the primary or secondary site based on the response it gets from the status file. To make this happen, one can simply rename the file and that will make sure the traffic is redirected to the site that is active.

Oracle Wallet should contain the certificate information

This step is pretty standard and the same approach that needs to be adopted to implement SSL in the first place. In short, copy the cer file over to the destination environment. In Oracle Wallet, Open and import the certificate and save it. Also, if you have a load balanced setup we can simply copy the cwallet.sso and the ewallet.p12 file to Oracle\Middleware\user_projects\epmsystem1\httpConfig\ohs\config\OHS\ohs_component\keystores\default

Update the ssl.conf,mod_wl_ohs.conf and httpd.conf files

In a nutshell, make the entries match the production server setup.

On httpd.conf

include "${ORACLE_INSTANCE}/config/${COMPONENT_TYPE}/${COMPONENT_NAME}/admin.conf"
include "${ORACLE_INSTANCE}/config/${COMPONENT_TYPE}/${COMPONENT_NAME}/mod_wl_ohs.conf"

include "moduleconf/*.conf"
include "conf/ssl.conf"
Include "conf/epm_online_help.conf"
Include "conf/epm_rewrite_rules.conf"
Include "conf/epm.conf"
Include "conf/epm_online_help.conf"
Include "conf/epm_rewrite_rules.conf"
Include "conf/epm.conf"

On ssl.conf


NameVirtualHost *:443
   
        ServerName example.hyperion.com
        RewriteEngine On

Restart OHS with these changes populated. The redirection must happen now.


Tuesday, March 24, 2015

Calculation Manager Rules that are still on Planning don't show up on Calculation Manager


In my 11.1.2.3.500 setup, I recently faced an issue with Calculation Manager Rules that were present on Planning forms inside the Planning Application, but however, on expanding the Calculation Manager->Planning App->Plan Type->Rules, it did not display the rules inside. These rules were imported manually earlier and suddenly disappeared. I could not really find what caused it but here are the steps to recover the BR's from planning tables to calculation manager:

1. Access the appropriate Planning application repository and its table HSP_CALC_MGR_RULES

2. Create a new text file with following content


3. Copy the contents in the Body column of the HSP_CALC_MGR_RULES table and paste its contents between the tag in the file created in step 2.
4. Save file as XML file
5. Within the Calculation Manager navigate to “File” -> “Import…” to import the Business Rule into Calculation Manager .

Using this I could restore all the rules back.

Tuesday, March 17, 2015

"Error executing query: -1" When Running Financial Reporting (FR) Reports

When we performed the migration to 11.1.2.3 since it was a new install some of the reports failed with the error "Error executing query:-1" .In addition to this a lot of the missing values were also being generated as a result. To resolve this issue, We had applied a fix in the previous environment as this issue seemed to exist there as well. We found out that this was due to the MDX query setting on FR Server and it was set to the default 'True' value. To carry out the same, follow the steps below.

1. Navigate to C:\Oracle\Middleware\EPMSystem11R1\products\financialreporting\bin and launch FRConfig.cmd


In some instances, this utility doesn't open and throws a java error. To fix that, follow step 2 below.

2. Edit setInstance.cmd and add the
set EPM_ORACLE_INSTANCE=C:\Oracle\Middleware\user_projects\epmsystem1

3. Now, it opens the FRConfig.cmd. In the MBeans tab, expand the Financial Reporting tab and scroll to the EssbaseUseMDX setting.


4. Restart FR services. This should fix the issue. 

I had found that on a parallel environment since it had loopback adapters present for the VIP, It did not work out for me. In that case, launch the enterprise manager module and correct it from there.Click on System MBean Browser.

Navigate to the Application MBeans


Also a side note, PreviewUserPOV can be set to True, so that end users don't have to set their FR Preferences to On. We can make the change on FRStudio.cmd






Thursday, March 12, 2015

BI+ distributed setup "startup document does not exist " "GSM unable to acquire LSM"

We have a network deployment for Foundation and BI+ in our environment and the 2nd node when operated on its own stopped being accessible. RAF agent service kept going to "stopping" mode when stopped and not completely stopped. The RAF Agent service and the Foundation Service was set to the installer account.

I thought it will be a good idea to compile a few things I verified in the hopes that it will help anyone with some troubleshooting ideas.

Ensure to set HyS9RaFrameworkAgent_epmsystem1 service to start with Domain account instead of Local System on both RAF servers and also make sure the domain account has local administrator rights .

Before starting services on Server2 make sure server1 is up and running and Workspace is accessible

Verify that HAVA ID in EPM registry matched in css_provisioning_info table

1: Execute the following query against the Shared Services registry database:
2: select application_id from css_provisioning_info where lower(application_id) like 'hava%';
o note this application id, e.g. HAVA:0000012b3072bf63-0000-6717-0acc1bac

3: Run epmsys_registry.bat from \Oracle\Middleware\user_projects\epmsystem1\bin
Open EPM registry.xml and search for HAVA and compare the HAVA ID using the ID above

4: If there is a mismatch, edit the Shared Services Registry:
Go to EPM_ORACLE_INSTANCE/bin and run the following command:
epmsys_registry updateproperty RA_FRAMEWORK/@applicationId

Where applicationId is the id returned from the query above. For example, using the id above:

epmsys_registry updateproperty RA_FRAMEWORK/@applicationId HAVA: 0000012b3072bf63-0000-6717-0acc1bac

5. Restart all services on Server1 and ensure Shared Services and Workspace are accessible

6: Enable the Service Broker service and the GSM on the second server
- Log on to EPM Workspace on Server1
- Select Navigate, then Reporting and Analysis, and then Services.
- Right-click Reporting and Analysis for server2 , and then select Properties.
- On the Services tab, set GSM to Enabled (Yes) and Service Broker to Enable (Yes).

Reproduce the issue and attach the logs from these locations from both servers 1/2
- \Oracle\Middleware\user_projects\epmsystem1\diagnostics\logs\ReportingAnalysis

and run epmsys_registry.bat from \Oracle\Middleware\user_projects\epmsystem1\bin
For my issue however, everything seemed to tie up quite nicely and I was still running into the issue.
This was the log entry from the server 2  raframework_stdout_console.log when started on its own:

Finally, when I started digging a little deeper, I could understand what was causing this to happen.
While Configuring the RAF for the 2nd node, We normally point to the shared server 1 RM1 directory while configuring.


The second Workspace agent was not able to locate the RM1 directory on server 1 .An issue with DNS could not resolve the server 1 hostname correctly.On second Workspace Agent server, add the server 1 hostname/IP address into the HOSTS file

On server 2, update the hosts file (which is under C:/Windows/System32/etc/drivers/hosts folder) with the ipaddress and the hostname of server1 where RM1 directory stored.
xx.xx.xx.xx server1.domain.com

This helped in resolving the issue.

.

Monday, March 9, 2015

Merge Cells Issue on Workspace Report

We have users that run the reporting out of Workspace and then export into Excel. Historically , The reports were formatted the same between Workspace and Smart View so you could easily pull from either and end up with consistent reporting.

Since the upgrade to 11.1.2.3, It appeared the new Workspace reports are no longer formatted the same. They had merged cells.


To fix the issue,

1. On your FR server, navigate to Oracle\Middleware\user_projects\epmsystem1\FinancialReporting\lib\FOProcessor
2. Save a copy of the existing xdo.cfg file.
3. Open xdo.cfg in a text editor.
4. Toward the top of the file, locate this line:




5. Immediately below it (above the
tag), add this line:


6. Save the file.
7. Restart the FR Web service (wait a few minutes to allow the service to fully render).

Wednesday, February 25, 2015

Calculation Manager Rules Runtime Tuning

We have had a case in 11.1.2.3 Calculation Manager where I have been trying to run the business rule, running them at the Total level in the 11.1.2.3 and I had to kill the calc ran after it ran for 5 hours.

In the process of debugging,I tried at a lower level to see if at least I can get it to complete. I did test the calc at a single intersection and it ran fine in seconds but that really wasn't much good to us
as our users all enter multiple intersections through their ad hoc forms and then need to calc at an aggregate level.

While starting to dig the rootcause , I thought it might be a good idea to share what was done to tune this rule. I started with the Cube cache settings.I have shown the values above.This needs to be followed with a Essbase restart.
Also for the application, I added these two properties . In order to do this, I logged into the planning application, Administration->Application->Properties. There, I added two properties called HBR_MAX_WAIT_FOR_RULE and HBR_MONITOR_DELAY and set them to 480000 which is 8 minutes.

Post this, I carried out a complete system restart and then ran the rule. Now what took a long time was completed in 5 minutes which is a big win. 

Wednesday, February 18, 2015

Hyperion SmartView Tuning

While driving the 11.1.2.2 to 11.1.2.3 upgrade in my organization, I have run into various problems to which some of my earlier posts are a testament.  Today, I bring to you one from the SmartView trove. We have users who swear by Excel Addin and refuse to let go of it. There have been some instances where users are unable to retrieve large numbers on SmartView.
This was the message that was produced when the retrieve was run on private connection using 19000 url .So, It would be something like, http://ohs.servername.com:19000/aps/SmartView .


 On using the SSL link for the same, This was the error message I encountered,

                                                                                                                                                                  
Listed below are a few ways in which the tuning can help in resolving the issue. First , I would tackle the browser timeout settings by the steps shown below.

On the client machine, update/add the following registry keys:

1. Open the Registry, Start -> Run -> Regedit.

2. Locate the following section(s):

[HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Internet Settings]\

or

[HKEY_LOCAL_MACHINE\Software\Microsoft\Windows\CurrentVersion\Internet Settings]

3.Create the following new DWORD keys with decimal values:

ReceiveTimeout=480000
KeepAliveTimeout=180000
ServerInfoTimeout=180000

In this example, the ReceiveTimeout setting is 8 minutes. The KeepAliveTimeout and ServerInfoTimeout settings are 3 minutes. Set these to greater than the longest running request.

4. Restart the machine for the new settings to take effect.

This in my case did not seem to help. There are some more tuning parameters we can add to aps.properties file located at C:\Oracle\Middleware\EPMSystem11R1\products\Essbase\aps\bin

# Specifies values for Essbase Server network connection
# APS will try netConnectRetry times to connect to an Essbase Server.
olap.server.netRetryCount=6000
olap.server.netConnectRetry=30
olap.server.netDelay=2000
olap.server.netSocketTimeOut=20000
Also, these values need to be added to the C:\Oracle\Middleware\EPMSystem11R1\common\EssbaseJavaAPI\11.1.2.0\bin\essbase.properties  file and try making a private connection to  http://aps.servername.com:13080/aps/SmartView
Adding these values helped me resolve the large retrieve issue. This bodes well with IE9 in the case of 11.1.2.3 .

Another request that is normally made is to increase the max rows to 65000 . In order to make that change, modify the properties shown below.

# Specifies maximum rows/columns for dataquery grid. 0 value indicates no limit.
service.olap.dataQuery.grid.maxRows=65000
service.olap.dataQuery.grid.maxColumns=255
To complete the change, login to EAS and expand the provider server tree. Right click on the provider server and edit the properties .


Restart APS . That Resolves the increase row limit Issue.

Tuesday, January 20, 2015

An unexpected error has occured while migrating outlines on EAS

I have been trying to migrate Essbase Outlines on 11.1.2.3.503 and have noticed this error some times on EAS. 

On looking on the EASServer.log there is nothing related to these errors and the messages panel below the EAS console doesn't display any error messages either. 

I have noticed that this occurs on two instances:

1. When you are not "connected" to Server1 or Server 2 Explicitly on EAS , i.e. when server 1 or 2 is not expanded .

2. When the target server is lower than 11.1.2.3 and the source server is 11.1.2.3

In my case, I could continually reproduce the error, when I wasn't connected to either of the servers and it went away after expanding server1 and server2.