Readme File for IBM® Spectrum Symphony 7.1.2 Cumulative Patch 441627

Readme file for: IBM Spectrum Symphony
Product/Component Release: 7.1.2
Update Name: Interim Fix 441627
Fix ID: sym-7.1.2-build441627-jpmc
Publication date: March 17, 2017

This cumulative patch provides various fixes for IBM Spectrum Symphony 7.1.2. For a detailed list of what’s included in this patch, see List of fixes.  

1.    Download location

Download cumulative patch 441627 from the following location: http://www.ibm.com/eserver/support/fixes/ 

2.    Installation and configuration

System requirements

Linux x86_64 

Before installation 

1.     Log on to the master host as the cluster administrator and shut down the cluster:

>  soamcontrol app disable all

> egosh service stop all

> egosh ego shutdown all

2.     If your cluster has only IBM Spectrum Symphony 7.1.2 installed, on every host in the cluster, back up the following files:

NOTE: Do not back up the perf_soam_loader.jar file in the $EGO_TOP/perf/soam/7.1.2/lib/ folder.

$EGO_TOP/3.4/linux-x86_64/etc/pem

$EGO_TOP/3.4/linux-x86_64/etc/lim

$EGO_TOP/soam/7.1.2/linux-x86_64/etc/sd

$EGO_TOP/soam/7.1.2/linux-x86_64/etc/ssm

$EGO_TOP/soam/7.1.2/linux-x86_64/etc/sim

$EGO_TOP/soam/mapreduce/7.1.2/linux-x86_64/lib64/libVEMApiCommon.so

$EGO_TOP/3.4/linux-x86_64/lib/jni/libVEMApiCommon.so

$EGO_TOP/soam/mapreduce/7.1.2/linux-x86_64/lib/libVEMApiCommon.so

$EGO_TOP/gui/3.4/lib/egogui.jar

$EGO_TOP/gui/3.4/lib/soamgui.jar

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/soamgui/servicePackage/addPackage.jsp

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/soamgui/js/wizard.js

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/symgui/sym/application/js/resourceList.js

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/symgui/sym/application/resourceList.jsp

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/soamgui/WEB-INF/classes/com/platform/soam/gui/api/SOAMResourceList.class

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/soamgui/WEB-INF/classes/com/platform/soam/gui/api/RealSOAMApiWrapper.class

$EGO_TOP/wlp/usr/servers/gui/apps/ego/3.4/platform/servicePackage/addPackage.jsp

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/soamgui/WEB-INF/classes/com/platform/soam/gui/api/SOAMAppList.class

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/soamgui/WEB-INF/classes/com/platform/soam/gui/api/SOAMAppConfList.class

$EGO_TOP/wlp/usr/servers/gui/apps/soam/7.1.2/symgui/WEB-INF/lib/soamgui.jar

$EGO_TOP/wlp/usr/servers/gui/apps/3.4/common_ui/js/common.js

$EGO_TOP/perf/soam/7.1.2/lib/perf_soam_loader.jar

$EGO_TOP/gui/conf/log4j.properties

 

If your cluster has multiple products installed (specifically IBM Spectrum Symphony 7.1.2, IBM Conductor with Spark 2.2.0, and IBM Spectrum LSF 10.1.0.0), on every host in the cluster, back up the following files in addition to the preceding files:

$EGO_TOP/wlp/usr/servers/gui/apps/3.5/common_ui/js/common.js

$EGO_TOP/gui/3.5/lib/soamgui.jar

$EGO_TOP/3.5/linux-x86_64/etc/lim

Installation 

1.       Log on to each host in the cluster and decompress the correct tar.gz file to your IBM Spectrum Symphony installation directory:

o   If your cluster only has IBM Spectrum Symphony 7.1.2 installed, decompress the sym-7.1.2.0_x86_64_build441627.tar.gz package to your IBM Spectrum Symphony installation directory:

> tar zxfo sym-7.1.2.0_x86_64_build441627.tar.gz -C $EGO_TOP

o   If your cluster has IBM Spectrum Symphony 7.1.2, IBM Conductor with Spark 2.2.0, and IBM Spectrum LSF 10.1.0.0 installed, decompress the symmh-7.1.2.0_ego3.5.0.0_x86_64_build441627.tar.gz package to your IBM Spectrum Symphony installation directory:

> tar zxfo symmh-7.1.2.0_ego3.5.0.0_x86_64_build441627.tar.gz -C $EGO_TOP

2.       Set the EGO_MANDATORY_DIR_CHECK_INTERVAL parameter in the ego.conf file to specify the time interval to check IBM Spectrum Scale status. If this parameter is not configured, the default value of 5 seconds is used.

For example, to set the check interval to be every 10 seconds, configure the parameter as follows:

EGO_MANDATORY_DIR_CHECK_INTERVAL=10

3.       Clean up the gui/work directories and clear the browser cache. Delete all subdirectories and files in the following directories:

NOTE: If you configured the WLP_OUTPUT_DIR parameter and set APPEND_HOSTNAME_TO_WLP_OUTPUT_DIR to true in the $EGO_CONFDIR/wlp.conf file, clean up the $WLP_OUTPUT_DIR/WEBGUI_hostname/gui/workarea/* directory. If you configured the WLP_OUTPUT_DIR parameter and set APPEND_HOSTNAME_TO_WLP_OUTPUT_DIR to false in the $EGO_CONFDIR/wlp.conf file, clean up the $WLP_OUTPUT_DIR/gui/workarea/* directory.

> rm -rf $EGO_TOP/gui/work/*

> rm -rf $EGO_TOP/gui/workarea/*

4.      If you configure a shared directory, on a management host, copy the log4j.properties file to the shared configuration directory:

NOTE: Before copying the file, back up the shared log4j.properties file.

> cp $EGO_TOP/gui/conf/log4j.properties $EGO_CONFDIR/../../gui/conf

5.      On each management host, remove the following unused file:

> rm -f $EGO_TOP/gui/3.4/lib/log4j-1.2.14.jar

In an environment with multiple products, also remove the following file:

> rm -f $EGO_TOP/gui/3.5/lib/log4j-1.2.14.jar

After installation 

1.     Verify the installation. If your cluster has only IBM Spectrum Symphony 7.1.2 installed, you might see output similar to the following:

> lim -V

  EGO 3.4.0 build 441627, Feb 14 2017

  binary type: linux-x86_64

> pem -V

  EGO 3.4.0 build 441627, Feb 13 2017

  binary type: linux-x86_64

> sd -V

  IBM Spectrum Symphony 7.1.2 build 441627, Feb 13 2017

> ssm -V

  IBM Spectrum Symphony 7.1.2 build 441627, Feb 13 2017

> sim -V

  IBM Spectrum Symphony 7.1.2 build 441627, Feb 13 2017

If your cluster includes multiple products, you might see output similar to the following:

> lim -V
  EGO 3.5.0 build 441627, Mar 10 2017
  binary type: linux-x86_64
  notes:
  fixes:

> sd -V

  IBM Spectrum Symphony 7.1.2 build 441627, Feb 13 2017

> ssm -V

  IBM Spectrum Symphony 7.1.2 build 441627, Feb 13 2017

> sim -V

  IBM Spectrum Symphony 7.1.2 build 441627, Feb 13 2017

2.     Log on to the master host as the cluster administrator and start the cluster:

> egosh ego start all

> soamcontrol app enable <appName>

Uninstallation 

1.     Log on to the master host as the cluster administrator and stop the cluster:

> soamcontrol app disable all

> egosh service stop all

> egosh ego shutdown all

2.     On all hosts, restore the backup files.

3.     Remove the EGO_MANDATORY_DIR_CHECK_INTERVAL parameter configuration.

4.     Clean up the gui/work directories. Delete all subdirectories and files in the following directories:

NOTE: If you configured the WLP_OUTPUT_DIR parameter and set APPEND_HOSTNAME_TO_WLP_OUTPUT_DIR to true in the $EGO_CONFDIR/wlp.conf file, clean up the $WLP_OUTPUT_DIR/WEBGUI_hostname/gui/workarea/* directory. If you configured the WLP_OUTPUT_DIR parameter and set APPEND_HOSTNAME_TO_WLP_OUTPUT_DIR to false in the $EGO_CONFDIR/wlp.conf file, clean up the $WLP_OUTPUT_DIR/gui/workarea/* directory.

> rm -rf $EGO_TOP/gui/work/*

> rm -rf $EGO_TOP/gui/workarea/*

5.     Clear your browser’s cache.

6.     Log on to the master host as the cluster administrator and start the cluster:

> egosh ego start all

> soamcontrol app enable <appName>

3.    List of fixes

408158 P101742

Summary

EGO fails to detect that a container in ZOMBIE state is older than the host uptime.

 

Description

EGO sends a kill -9 command to an old service PID when it fails to detect that a container in ZOMBIE state is older than the host uptime. Apply this fix to avoid wrongly killing the old service PID.

 

Symptoms

EGO sends a kill -9 command to an old service PID after the host is rebooted.

 

Impact

It's possible that sending kill -9 to old PIDs could kill a living unrelated process.

 

Component

PEM

 

409974

P101754

Summary

When SSM startup times out, it is killed twice, causing SSM to not restart.

 

Description

when SSM startup times out, applications cannot execute because SSM is restarted by two threads simultaneously.

 

Symptoms

SSM is restarted twice by two threads almost at the same time.

 

Impact

Applications cannot run normally.

 

Component

SSM

 

414087

P101801

Summary

When using the subscribeTaskProgress API and you submit two sessions, then kill one session, SIM restarts.

 

Description

When using the subscribeTaskProgress API in Platform Symphony Developer Edition or Platform Symphony, if you submit two sessions and then kill one, SIM restarts. This causes the remaining session to re-run.

 

Symptoms

SIM restarts and workload is re-run.

 

Impact

Workload re-runs.

 

Component

SSM

 

416064 P101823

Summary

A package with a dash (-) in the name cannot be uploaded using the management console

 

Description

Uploading a package using the management console fails when the package name contains a dash (-), such as abc-1.zip.

 

Symptoms

Fail to upload a package with a dash in the name.

 

Impact

Cannot upload a package with a dash in the name.

 

Component

PMC

 

418281 P101868

Summary

The management console memory continually increases until it reaches the MEM_HIGH_MARK value.

 

Description

Due to constant page refreshes, the management console requires increased memory. This occupies so much memory that the management console reaches the maximum memory mark, which triggers Platform Symphony to restart.

 

Symptoms

Memory occupation increase

 

Impact

Management console malfunctions

 

Component

PMC

 

418666

P101861

Summary

Unable to select a host name from application resources list page.

 

Description

When you view resources for an application from the management console (at Workload > Symphony > Applications, selecting an application to view and then selecting the Resources tab), you see only the host name and are unable to select it.

 

Symptoms

Host names are not selectable on the resource list page.

 

Impact

Host names are not selectable on the resource list page.

 

Component

PMC

 

419908

P101879

Summary

Providing a non-existing host name on the host list page shows an error message. Then, switching to the host icon page, shows an exception string instead of a message.

 

Description

In the management console, when you access Resources > Hosts > Allocated Hosts > Host (List View) to view a list of hosts, provide a non-existing host name, and click Find, an error correctly displays. When you switch to the Hosts (Icon View) page, the console throws an exception string, without a usable error message.

 

Symptoms

Host list page throws an exception string without a usable error message.

 

Impact

Host list page throws an exception string without a usable error message.

 

Component

PMC

 

420226

P101880

Summary

When there is a <appname>_session.soamdb.<timestamp> file generated in the $SOAM_HOME/work/history/ directory, the session history loader reads this file and delays for several hours, causing missing data.

 

Description

When there is a <appname>_session.soamdb.<timestamp> file generated in the $SOAM_HOME/work/history/ directory, the session history loader delays for several hours to read this file and causes missing data. The session history loader log file shows many errors regarding unique constraint violations.

 

Symptoms

Many errors regarding unique constraint violations.

 

Impact

Data lost in the database.

 

424073

P101926

Summary

You are unable to use the management console to update a service package associated with an application profile.

 

Description

While updating an existing service package that is associated with an application profile, the management console throws an exception regarding the use of multiple points.

 

Symptoms

Update a service package throws an exception of multiple points.

 

Impact

Cannot update a service package from the management console.

 

Component

PMC

 

426239

P101950

Summary

Upon experiencing a high volume of traffic, system traffic returns to normal, but LIM still shows that hosts are in busy state.

 

Description

If the system experiences a high volume of traffic, this can lead to /proc/stat operation failure. However, the system will continually show a system busy state, even when system traffic has returned to normal. LIM does not correctly report system load status. 

 

Symptoms

LIM incorrectly reports hosts are in busy state.

 

Impact

LIM incorrectly reports hosts are in busy state.

 

Component

LIM

 

422392

P101908

Summary

You are unable to sort by the Slots column on the Resources tab within the management console.

 

Description

You cannot sort using the Slots column on the Workload > Symphony > Applications > Resources tab on the management console. You are unable to sort by ascending or descending order.

 

Symptoms

Slots column cannot be sorted.

 

Impact

Slots column cannot be sorted successfully.

 

Component

PMC

 

420751

P101885

Summary

For a Platform Symphony 5.1 or 6.1.1 environment, when you submit workload using the 5.1 or 6.1.1 client to a Platform Symphony 7.1 cluster, and a task fails, the client shows an ECODE_ERROR_NOT_PROPAGATED_BUT_EXPECTED exception.

 

Description

In an environment where the client and service are compiled by Platform Symphony 5.1 or 6.1.1, and you use the 5.1 or 6.1.1 client environment to submit workload to a 7.1 cluster, when a task fails, the client encounters the following error, which is wrong and misleading:

 

“Domain <SOAM>: Internal Error: ECODE_ERROR_NOT_PROPAGATED_BUT_EXPECTED.

The exception associated with the task result was not propagated to the client. Contact Platform technical support.”

 

This fix ensures that SSM sends the proper error message to the client.

 

Symptoms

Misleading error message to 5.1 or 6.1.1 client

 

Impact

Platform Symphony 5.1 or 6.1.1 application cannot be run on IBM Spectrum Symphony 7.1.2

 

Component

SSM

 

408431 P101740

Summary

When IBM Spectrum Scale is used for high availability (HA), LIM will hang when the IBM Spectrum Scale status checking takes a long time due to IBM Spectrum Scale being slow or inaccessible.

 

Description

When IBM Spectrum Scale is used for HA (with EGO_MANDATORY_DIR configured), LIM will check the IBM Spectrum Scale’s status periodically. When LIM checks, it will wait for the check result until a successful or failed result, which may make LIM unresponsive for a period of time. 

 

With this fix, LIM will check the IBM Spectrum Scale’s status without blocking the system, and LIM can still handle requests normally. The last master candidate will not exit even if the IBM Spectrum Scale failure time exceeds the defined time period.

 

Symptoms

LIM unresponsive on unstable IBM Spectrum Scale.

 

Impact

LIM unresponsive

 

Component

LIM

 

427720

P101970

Summary

Error message “Too many users are accessing the Console at the same time. Try again later.” displays in the cluster management console.

 

Description

Currently, if you enable or disable applications within the cluster management console more than 30 times, the following error message displays: “Too many users are accessing the Console at the same time. Try again later.”

 

After applying this fix, the error no longer displays in the cluster management console.

 

Symptoms

Error message displays and some SOAM related pages cannot be used.

 

Impact

User can't use cluster management console to manage applications

 

Component

PMC

 

433117 P102017

Summary

When the <appname>_session.soamdb.<timestamp> file is generated in the $SOAM_HOME/work/history/ directory, the session history loader delays for several hours, causing missing data.

 

Description

After applying this fix, no data is lost.

 

Symptoms

The session history loader delays for several hours to read the <appname>_session.soamdb.<timestamp> file and causes missing data. 

 

Impact

Missing data.

 

Component

PERF

 

441269

P102096

Summary

SIM encounters errors from SI methods “Register” and “onSessionEnter”, triggering error handling for both methods.

 

Description

In some cases, when the Register method fails, error handling is triggered for both the Register and onSessionEnter methods. This fix resolves the issue, so that error handling is triggered only for the Register method

 

Symptoms

Error handling for both “Register” and “onSessionEnter” methods are triggered.

 

Impact

Unexpected error handling.

 

Component

SIM

 

441251

P102093

Summary

Unresponsive menu in cluster management console

 

Description

This interim fix resolves unresponsive menu issues when using the cluster management console in IBM Spectrum Symphony 7.1.2.

 

Symptoms

Menu items become unresponsive periodically.

 

Impact

Unable to perform cluster management operations from the cluster management console.

 

Component

PMC

4.    Copyright and trademark information

© Copyright IBM Corporation 2017

U.S. Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.

IBM®, the IBM logo, and ibm.com® are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the Web at "Copyright and trademark information" at www.ibm.com/legal/copytrade.shtml.