Known problems, limitations, and workarounds

ID

QCCR1D234930

Problem

Cloud Optimizer is not able to get CSA Organization LDAP details after fixing existing security vulnerability in HPE CSA

Cause

Code fix is completed to close the security vulnerability. Please ensure to pass the userIdenttifier to all the API calls as mentioned in CSA REST API Guide.

Workaround

Hotfix available. Please contact support.

ID

QCCR1D228812

Problem

Codar 1.70.RC3. Privacy Statement is missing in Codar Login Page. The privacy statement does appear to show up on 4.8MR bits on the initial MPP login screen at /org/<ORG_NAME>.  However, the next IDM screen you are navigated to when clicking "Log in" is where the issue lies. The screen apparently should, but does not show the privacy statement. 

Cause

Code defect

Workaround

The user could always manually go to the privacy link they have configured for the organizations privacy statement. This link is available in the Organizations > Organization Name > General Settings area.  Or, the user could simply return back to the main portal login page to view the privacy link.

ID

QCCR1D235507

Problem

Resuming a failed upgrade is failing on Oracle PCoE Environment because of the created_by and updated_by references in the scripts.

Cause

Resuming a failed upgrade is failing and csa\_CSA_4_80_0_installation\Logs\install.log contains error message that UPDATED_BY_ID columns is missing  

For example error message from Oracle database:

PL/SQL: ORA-00904: "UPDATED_BY_ID": invalid identifier

Workaround

Restore the database from backup taken prior upgrade and resume the upgrade again.

ID

QCCR1D227860

Problem

UserGroup indexing to Elasticsearch REST call throws an exception even though the indexing is fine. Exception in log file, csa.log

Cause

A bug in NodeJS.

See https://github.com/nodejs/node/issues/712

Workaround

No workaround is required as this does not influence the search behavior and MPP full text search works.

ID

QCCR1D229293

Problem

Operation Orchestration displays: 403 Error when  trying to use direct link from HPE CSA using menu Workflows -> Orchestration in case that  LDAP Users does not Exist in OO 

Cause

Caused by defect  QCCR8C32250 in product HPE Operation Orchestration

Workaround

Modify URL of HPE Operation Orchestration to end with  /oo/login/login-form

or use seeded user admin for login to HPE CSA.

ID

QCCR1D232747

Problem

SSO - session not invalidated while logout/login as different user in OO. When CSA is configured with HPSSO which is a default in 4.8 (documentation contains steps how to disable it for FIPS mode), it will fail to logout user from CSA when the user performs logout action in other integrated product like HPE OO followed by login as different user in HPE OO.

HPSSO does not have central server to handle logout. Programs using HPSSO do not need to know each other as long as shared secret used for encryption and validation of cookie is same.

Cause

In CSA the Management Console in csa.war uses IdM tokens as means to authenticate users. HPSSO tokens are recognized only by Identity Management (IdM) component, but not by CSA Management Console. When there is no session with Management console, the csa.war will redirect user to IdM which will validate and consume HPSSO cookie to produce IdM token for Management Console.

CSA Management Console will perform logout when it is accessed while the HPSSO cookie is missing (due to logout in other product which deletes the cookie).

But if CSA Management Console is accessed after logout followed by login in other product, then it will not detect the logout (it does not inspect cookie contents, it cannot decrypt it, in CSA only IdM component can decrypt it) and keep session established with original user identified by IdM token.

Sessions are still limited by token expiration. IdM tokens expire after 30 minutes by default, though there is process to extend their lifetime if done within these 30 minutes.

Workaround

Multiple options:

- Turn off HPSSO if its functionality is not desirable.

- Perform logout in CSA, not in HP E OO.

- After performing logout in HPE OO, access HPE CSA so HPE CSA notices the cookie got deleted and it will perform logout, before continuing in to login as different user in HPE OO.

ID

QCCR1D235119

Problem

SQL statement execution timeout cannot be applied to install/upgrade scripts.Upgrade fails with "Read timed out" during processing "csa_remove_createdby_updatedby.plsql" file on SQL Server (check CSA\_CSA_4_80_0_installation\Logs\install.log to ensure it)

Cause

csa_remove_createdby_updatedby.plsql contains ALTER TABLE ... REBUILD statements that goes through all rows in table. So the REBUILD operation could timeout on huge databases.

Workaround

1) Edit CSA\scripts\mssql\csa_remove_createdby_updatedby.plsql and comment out all "ALTER TABLE ... REBUILD" statements at the end of the file. You can comment it out by enclosing all REBUILD statements to /* and */ (SQL comments). Save the file.

   2) Restart the upgrade process. It should end successfully.

   3) After upgrade it is recommended to apply the commented REBUILD statements manually, For example, using Microsoft SQL Server Management Studio.

ID

QCCR1D235209

Problem

Semantic Validation for properties in User Operations->Public Action fails to recognize the variable. If [CLIENT:<prop>] token is used as parameter value of semantic validation parameters, then test (validation) fails with following error message:

Option property with the name <prop> is not found. If this refers to a new property, ensure that it is saved.

Test fails even if the [CLIENT:<prop>] is used on property <prop> and user filled "Input Validation" input fields.

Cause

1) if token [CLIENT:<prop>] is used as parameter value of  validation parameter  on user operation parameter <prop>

      and user filled Input Validation field input

   -> it is caused by UI, which does not replaces the [CLIENT:<prop>] token with string entered to Input Validation input field.

2) if token [CLIENT:<prop>] refers to another parameter of user operation

  -> it is caused by backend, which does not resolve tokens pointing to parameters of user operation when test is requested

3) if token [CLIENT:<prop>] refers option model property

  -> such token cannot be resolved because

       a) user operation does not have access to option model (e.g., user operation is on resource offering),

       b) option model can define more than one property of the name, so it is not possible to identify the property

Workaround

For all causes 1), 2), and 3) the workaround is

* enter value instead of  token [CLIENT:<prop>] used as user operation parameter value

  - once tested, change parameter value back to the token

ID

QCCR1D234010

Problem

CSA 3.2 to 4.8 upgrade path-   the increment and decrement quantity field in shopping cart is not showing or working properly. The increment and decrement quantity field in Marketplace Portal shopping cart does not show increment/decrement buttons in Internet Explorer.

Cause

Standard UI widget for number input type does not include increment/decrement buttons in Internet Explorer.

Workaround

Enter quantity manually.

ID

QCCR1D235356

Problem

CSA 4.8 Content Store, after installing the content from file system, if we try installing the same content through HPLN, it does not prompt for re-install dialog box in the first attempt.

Cause

Product Limitation

Workaround

1. Install the content from file system.

2. Install the same content from the HPLN site (it will fail for the first time).

3. Re-attempt because all subsequent attempts to install the same content from HPLN site will succeed and the re-install dialog box will appear.

Refresh the browser after every attempt in case the browser is slow or lagging.

ID

QCCR1D235590

Problem

Import Preview fails while previewing the import of a design from an upgraded CSA instance into a fresh CSA 4.8 instance.

Cause

A constraint violation is indicated in the Preview

Workaround

Result of Import Preview can be ignored, and the design can be imported by clicking on Import button instead.

ID

QCCR1D234938

Problem

CO URL in csa.war/dashboard/config.json file  has changed to default value after Upgrade to CSA 4.8.

Cause

Structural changes to the config.json file that is the New Dashboard changes

Workaround

If you have made customizations to the config.json for the CSA launchpad/dashboard you will need to manually re-apply those after upgrade due to structural changes to the config.json file with the addition of the new interactive dashboard.

ID

QCCR1D234038

Problem

SAML configuration is lost after upgrade to CSA 4.8 from CSA 4.7

Cause

Process Limitation

Workaround

If SAML is configured in CSA 4.7 and you have upgraded to CSA 4.8, you need to configure SAML again.

ID

QCCR1D225958

Problem

Cloud Optimizer (CO) Dependency: When the VM is powered OFF, CO is not sending the ideal data points (zero values) to be displayed for Memory and CPU graph in Health tab. Missing data points when VM is powered OFF or Suspended.

Cause

Unable to plot the graph for missing data points.

Workaround

No workaround available.

ID

QCCR1D228220

Problem

Health status is not updated for servers deployed on Helion Openstack (HOS) provider.

Cause

CSA is unable to retrieve the health status since Cloud Optimizer (CO) is not supporting HOS 3.0. 

Workaround

It is a product limitation. No workaround available.

ID

QCCR1D228619

Problem

Global search/Elastic search from MPP portal does not work in a Linux CSA installation

Cause

CSA Search service fails to update the Elasticsearch indices as a result of which Global search from MPP returns nothing

Workaround

After CSA installation is complete, or after a CSA restart, stop the CSA Search service and restart it manually by following the steps below:

If CSA was installed in a location other than  /usr/local/hp/csa, adjust the path accordingly.

ID

QCCR1D234418

Problem

IdM Dependency - Upgrade from 4.6 to 4.8 is failing for windows with MS-SQL database.

Cause

Snapshot Isolation not enabled for idm

Workaround

For Microsoft SQL Server, it is mandatory to enable the snapshot isolation for Identity management database which can be achieved through following two database statements:

ALTER DATABASE idmdb SET ALLOW_SNAPSHOT_ISOLATION ON;

ALTER DATABASE idmdb SET READ_COMMITTED_SNAPSHOT ON;

ID

QCCR1D218883

Problem

Custom changes in Elasticsearch configuration may be discarded during an HA upgrade installation. Elasticsearch configuration file resets and custom changes in the elasticsearch.yml file are lost

Cause

Product defect.

Workaround

Custom changes from upgraded installation are stored in a backup folder in /elasticsearch/config/. Transfer custom changes from the older installation file into the upgraded file.

ID

QCCR1D227598

Problem

The SAML Authorization does not work for the CSA Provider, if the access control is configured with the LDAP sub tree at the Organization Unit (OU) Level.

Cause

CSA does not support the LDAP sub tree for Access Control (ACL) when SAML is enabled.

Workaround

None

ID

QCCR1D235063

Problem

In Operations UI - “Subscription Status", "Service Instance Status", and "Upgradable To" fields go out of focus and get shifted to the bottom of the Operations Overview page.

Cause

This is a Chrome browser issue in version 52 through version 55. 

Workaround

Use Chrome 56 and above version to avoid this problem.

ID

QCCR1D234562

Problem

SMC login fails if we disable HP SSO configuration manually on CSA 4.8 HA

Cause

Product limitation

Workaround

Open the file $CSA-HOME\jboss-as\standalone\deployments\csa.war\WEB-INF\applicationContext-security.xml and set checkSSOCookie value to false in the below mentioned section of bean and restart CSA service.

<beans:bean id="tokenValidityFilter"class="com.hp.csa.security.TokenValidityFilter">

<beans:property name="checkSSOCookie" value="true"/>

</beans:bean>

ID

QCCR1D230155

Problem

IDM Dependency - HPSSO: SMC portal does not get logged out upon HP SSO timeout.

Cause

Product limitation

Workaround

No workaround currently available

ID

QCCR1D230605

Problem

Manage User Subscriptions under Administration of MPP lists few users unauthorized to access the MPP

Cause

Product limitation

Workaround

No workaround currently available

ID

QCCR1D235314

Problem

Even if the undeploy/unreserving actions fails during cancellation, the subscription will go offline. However the resources may not be completely deallocated.

Cause

This is because of a limitation in the architecture that does not allow to pause/resume the failed actions during cancellation.

Workaround

No workaround available

ID

QCCR1D232661

Problem

Currently we do not support hybrid Cloud Slang-AFL flows in any combination.

Cause

It is a product limitation.

Workaround

The procedure on ‘Creating a topology Design with CloudSlang’ is available in the Topology Components Guide (Whitepaper). Refer to this guide for a workaround information.

ID

QCCR1D186068

Problem

Imported topology component does not present a list of values for a multi-select Input property. When importing an Operations Orchestration flow in the Designs / Topology / Components area of the Cloud Service Management Console, if that flow contains an Input property with Type value of List of Values and From value of Prompt User from List – Selection List, the resulting component imported into CSA will have a property value of type String for this Input property. Instead of a list of values from which one or more can be selected, a single text input will be presented to the user for this property in both the Components and Designer areas.

Cause

The Designs / Topology / Components and Designs / Topology / Designer areas of the Cloud Service Management Console do not have graceful support for multi-select properties such as these.

Workaround

In the text input for such a property, encode the property values using the appropriate delimiter, which is determined by the method the flow uses to parse the Input property. If the flow uses the Selection List Iterate operation that is provided with the Base content pack in Operations Orchestration, the delimiter (separator) is configurable and has a default of ‘|’. For example, the values ‘red’, ‘green’, and ‘blue’ would be specified as ‘red|green|blue’ (unquoted) if using the Selection List Iterate operation with the default separator value.

ID

QCCR1D187711

Problem

Topology component Boolean and Integer properties are not correctly mapped to Chef Attributes. Topology components imported from Chef include an attributes parameter in their deploy operation, allowing customization of the provisioning of the Chef recipe. Properties passed in the attributes parameter are automatically converted to Strings. For example, an Integer component property of 3306 will be converted to "3306", and a Boolean component property of true will be converted to "true". If the Chef recipe is written to expect an Integer or Boolean input and not a String, the provisioning of the component will fail.

Cause

Product limitation.

Workaround

The Chef recipe should be written or modified to expect String inputs.

ID

QCCR1D226494

Problem

The Featured Category list is empty for a newly created organization

Cause

The organization data synchronization is not complete after a new organization is created in IDM tables.

Workaround

After the synchronization is completed, the catalogs and featured category list will appear. (~30 seconds)

ID

QCCR1D233354

Problem

MPP - Groups the user belongs to is not displayed during the submit and modify operations. In MPP Service checkout page for an offering, the group list shows only the DNs that are added in the access control of the organization, it does not list all the groups to which the user belongs in LDAP.

Cause

This behavior is currently unsupported in CSA. In order to show the group in group list, all the groups need to be explicitly mentioned in the organization access control.

Workaround

None

ID

QCCR1D234644

Problem

Misleading icon displayed in MPP.

If an organization is set to Pause Subscriptions on Provisioning error and a subscription fails, MPP shows right status as Paused but the icon is wrong.

Spinner is displayed in MPP instead of Pause icon.

Cause

Wrong HTML in MPP code base.

Workaround

No workaround is required. Inappropriate icon is displayed.

ID

QCCR1D228672

Problem

Cannot launch the show performance page using SSO from MPP.

Cause

SSO token is not passed correctly.

Workaround

User can login to Cloud Optimizer manually by entering username and password.