IBM Enterprise Content Management System Monitor Interim Fix ESM_5.5.5-000-001

The P8 Audittrail probe no longer fails with a NullPointer-Exception if a log entry contains null values in CLOBs.

It is now possible to disable the default behaviour to automatically deploy a basic monitoring to new agents. This has to be done by creating the text file <server install dir>/karaf/etc/de.cenit.phoenix.server.services.agents.AgentBasicMonitoringInitializer.cfg with the following text in one line as its only content: disable=true

A memory leak in the monitoring view in combination with Firefox was fixed. But please see the related Known Limitation for restrictions, that could not be fixed for now.

The sometimes observed behaviour of agents losing the connection to their server after some time is fixed.

Scheduled Reports are now saved on disk in the directory "<server install dir>/karaf/server/reports/" as PDFs instead of being saved in the database.

A continuously blocking task that should forward incidents no longer leads to block the sample processing.

The dashboard settings are now also included in an export and the probes are no longer deactivated. Instead the user now must explicitly disable the probes while importing.

IBM JRE specific JSSE2 system properties to enable TLSv1.2 are now always set.

The GeneralDiskspace probe now correctly handles continuation lines in df output.

The table of subsystem references in the configuration wizard now has descriptive column headers.

Cumulative security update for IBM Java 8 JRE included in the product.

Narrower title leaves more space for information bar, which also acts as a navigational breadcrumb in views that are showing a hierarchy.

It is no longer possible to define a schedule for callback probes (e.g. logfile probes). These do not need a schedule as they are triggered automatically. So the definition of a schedule has no effect et all.

If HTTPS is enabled accessing the web-UI through the HTTP-URL the user will automatically redirected to the HTTPS URL.

Self-monitoring: It is possible to automatically send emails if one or more agents are in the state of "timed out" and/or if one or more tasks failed to run properly (state "failed" and "canceled").

For ODM logfile monitoring the formatfile for the IBM WebSphere logfile can be used.

The CMOD probes will no longer fail if the CMOD installation is not in the default directory.

Custom v2s format files are now retained between updates of the server and the agent.

All file names and domain user names can now be entered in the correct OS specific format. This means that it is no longer required to replace backslashes in file names and user names by forward slashes.

An exported configuration can now be imported. The import reads a JSON file that was created through an export, see extra release notes for that feature.

Export of the configuration is now possible. In this first version one can now export the configuration, without the situations and probes to a JSON file. That file can later be used for an import, see extra release notes entry for that feature.

The WebResponseTime and WebStatusMonitor probes have been switched to a pure Java implementation. Please note that due to technical changes the FULL_RESPONSE_TIME option for the WebResponseTimer probe is no longer available. Probes that use this option will automatically be changed to the option TOTAL_DURATION instead. Also please note that the both probes currently don’t support NTLM authentication.

The LdapConnection and LdapPerformance probes have been switched to a pure Java implementation. Please note that the login.conf for existing probes must be adjusted. Refer to the Configuration Guide for an example of a login.conf file.

The CaseManagerCaseStatus, CaseManagerSolutionStatus and CaseManagerTaskStatus probes have been switched to a pure Java implementation.

The CustomDatabaseQuery probe no longer ignores its timeout setting. Note that the setting is now in seconds instead of milliseconds.

New agents connecting to the server the first time will get a set of probes for monitoring basic paramesters like CPU, memory, and diskspace. Existing agents will not be affected.

Deep copy of subsystems is now supported. A deep copy is a copy of a subsystem populated with copies of the references of the child subsystems found in the original.

Shallow copy of subsystems is now supported. A shallow copy is a copy of a subsystem populated with references of the child subsystems found in the original.

Cumulative security update of the IBM Java 8 JRE, shipped with the product.

The CMOD probes will no longer fail if the CMOD installation is not in the default directory.

To prevent irritations a status page is shown in the browser till the web-UI is ready for logons.

New probe to report the number of BPM processes in a specific state

It is now possible to define some more obscure expert options for SMTP in the SMTP server settings.

The listener license/usage probes could fail if a Listener is not sending any listener instance data, because the probes had handled such a listener as an older implementation. This could lead to an unspecified exception in the listener code. With this fix, the probes no longer support that older license/usage API of the listener. Only the license/usage data provided by the ILMT listener paths is now used regardless of the existence or non-existence of any instance information provided by the listener.

All non-vital ports of the agent and the server are now closed per default. Especially the H2 web- and database-server ports are closed. Although it was planned to close the web-server of the agent, the port is still open. That is because the Jetty of the agent is used now by the new logstash probe.

The message in samples created by the ResourceManagerHeartbeatProbe now correctly indicates if the LAN cache is on or off.

New probe to count the number of BPM process application snapshots.

SNMP task editor: It is no longer possible to close the editor of a newly created task without a name by clicking on the save button. The editor stays open with a warning. The editor can be closed without saving through clicking on the cancel button.

A newly created task is no longer erroneously made active, when saved. The user has to do this manually.

SMTP forwarding with security enabled (user/password needed for SMTP server access), works now as expected.

All file names and domain user names can now be entered in the correct OS specific format. This means that it is no longer required to replace backslashes in file names and user names by forward slashes.

The H2 database server is no longer accesible from outside the application..

After an update to 5.5.2.0-002 with multiple dashboards defined the dashboards are no longer overwritten with the content of the first dashboard.

Subsystem selection in task editor is now cleared if no matching subsystem is found. If the task editor was opened for a task type, where a subsystem of the corresponding type is defined and then the task type was changed to a type where no corresponding subsystem exists, the previously selected subsystem remained selected. This is no longer the case.

Active and inactive probes assigned to an agent are listed in the info/details section of an agent. The bug was, that there were only active probes listed.

The configuration settings and the monitoring data are now always stored in two separate databases. This allows the monitoring data to be stored in an external IBM DB2 database instead of the embedded H2 database. The configuration data is always stored in the embedded H2 database. It is not possible to select IBM DB2 as a database when updating. It is only possible to select IBM DB2 as a database when executing a (new) full install.

The following issue in the monitoring view is now fixed: In a multi-tab dashboard selecting one of the incidents in a graph portlet in a dashboard not the last one and without a scrollbar, will scroll-up the complete dashboard content. Parts of the dashboard are out of sight and below the tab row and header of the page. There is no scrollbar to refocus the dashboard.

Flexera’s InstallAnywhere 2020 is now used as the installer for the product.

A memory leak in the web ui was fixed. This memory leak could be triggered by repeatedly switching between the Situations and the Task Portlets icon on the outer most left side of the Monitoring view.

Fixed context root prevented access to webconsole e.g. via Ingres in AWS.

No longer needed fields of the CPE subsystem were removed. The fields are "installation path", "OS user", and "API installation path".

The agent counter in the status icon in the upper right corner of the web-ui now shows the correct amount of agents per status.

The subsystem configuration wizard now shows descriptive short labels below the subsystem icons to ease the identification of the underlying subsystems.

If a probe returns more than one sample in one run, each sample is forwarded into the evaluation, so for each of them an incident can be created. This was a known limitation, which no longer exists now.

The probes Icc4SapArchiveStatus and Icc4SapServerStatus failed on UNIX and Linux systems due to an incorrect directory name. This has been fixed.

Icc4SapArchiveStatus probe no longer fails when authentication type is selected or more than one archive is defined in the INI file.

Icc4SapProcess situation no longer fails with the error message The following processes do not have the specified number of instances […​].

The Listener probes have been converted to a pure Java implementation. The following fundamental changes have been implemented: * the probes now run inside of the agent JVM, no external JVMs are started * the probes no longer need a schedule; data is sent as soon as it is available * the configuration now consistently uses the path and column names from the IBM System Dashboard for Enterprise Content Management * there is no longer a distinction between USER and METER data: meters are now part of the USER branch as in the IBM System Dashboard for Enterprise Content Management; for this reason, the Ce*Meter* probes have been merged into the corresponding Ce*User* probes, existing probes are converted during update * due to the technical changes, the analyzation type has been removed from all Listener probes

New portlet to build a kind of a heatmap in a dashboard implemented.

Support to use the asterisk wildcard in logfile paths and names added.

Incident of the Port Reachable probe showed a timeout of zero milliseconds as January 1st, 1970 00:00:00 UTC. This is now fixed and zero is shown as the correct value.

Setting the languages in the browser setting to a language code alone, e.g. "en" or "de", but not setting the country code, e.g. "English (United States Of America)" or "Deutschland" prevented the web-UI from loading. Now the web-UI will be loaded correctly.

Security fix: It is now possible to turn the H2’s web server off by not setting its port in the installer. This works also while updating the product not only new installations.

A memory leak in the web ui was fixed. This memory leak could be triggered by repeatedly switching between the Situations and the Task Portlets icon on the outer most left side of the Monitoring view.

Filtering of samples on the Agent side works now reliable and can be configured through the Web-UI.

Selecting a task type and clicking on the add (plus) button now correctly opens the task editor for the selected type.

Exporting of incidents to slack works now.

Cumulative security fixes for (IBM Java 8) JREs included in the ESM 5.5 product line.

Some cepest based probes unnecessarily used an internal locking mechanism to prevent parallel runs of the same probe. This could led to errors if the probe’s execution time took too long compared to the schedule. These probes are no longer utilizing the internal locking mechanism.

Monitoring of rotating logfiles no longer unexpectedly stops. The monitoring of so called rotating or rolling logfiles stopped unexpectedly in cases where the logfile to monitor has got a new reference handle (aka I-node) in the file system when "rolling" occurred.

Fixed context root prevented access to webconsole e.g. via Ingres in AWS.

Security scans of the MQTT port of the server via OpenVAS no longer leads to high CPU loads.

Continuous update of all portlets works now reliable even after switching between "Monitoring" and "Configuration" back and forth.

All details of an agent shown in the sidebar below the list of agents are updated if the status of the corresponding agent changes.

New installations no longer use non-standard ports for HTTP and HTTPS for the web-ui.

The problems with the automatic start after the installation no longer exists. The issue was removed from the list of knonw limitations.

The probes Icc4SapArchiveStatus and Icc4SapServerStatus failed on UNIX and Linux systems due to an incorrect directory name. This has been fixed.

Icc4SapArchiveStatus probe no longer fails when authentication type is selected or more than one archive is defined in the INI file.

Icc4SapProcess situation no longer fails with the error message The following processes do not have the specified number of instances […​].

The WMI probe no longer triggers core dumps on agents.

The ICN plug-in now filterts the incidents by the selected situation group.

The limitation of columns in the monitoring dashboad does no longer exist.

The return value of the MSSQL Processes Probe was always 0. The correct return value was only shown in the message field. This is now fixed. The probe returns 0 (zero) and 1 (one) accordingly to measured state’s of the MS SQL processes.

All server’s and the agent’s TLS based communication now uses only Diffie-Hellman (DH) keys sizes bigger or equal to 2048. Shorter key lengths are no longer supported. So e.g. agents prior to this fix will not be able to connect to the server after it was updated. All agents must be updated, too, to allow them to reconnect to the server. This fix can also have effects on HTTPS based communications. It can also have an effect on the overall performance while starting the server. If you want to downgrade to a key length of 1024 after the update, you can edit the JVM property " jdk.tls.ephemeralDHKeySize" in karaf/etc/system.properties. After a change of this property, the server has to be restarted to apply the change to the JVM.

Microsoft Internet Explorer 11 only: Sometimes a tag element in the search dialog of the situation group list dispeared. Only manual refresh of the GUI showed the tag again. This behaviour is now fixed. The shown tags are stable now.

The incident list for a situation can now correctly be sorted by timestamp.

Listing or saving reports could lead to rest errors like "Server could not perform requested action". This could only be fixed by upgrading a third-party library.

The Listener probes have now access to all counters reported by the Listener API, especially the raw counters and accumulators and their statistics data.

Cumulative security update of the IBM Java 8 JRE, shipped with the product.

All probes for subsystems of type eDiscovery received a wrong number of arguments. This has been fixed.

The probes ContentNavigatorPingPageStatus and SyncServerPingPageStatus failed for ICN subsystems that contained a URL without port. This has been fixed.

The DB2Statistics probe now works for DB2 installations that use other languages than English.

The DB2Statistics probe failed because the GSK Kit was missing from the search path.

To streamline the LDAP configuration and to prevent confusion, the configuration settings were split of in two sections. A standard fields section, which is shown immediately after the LDAP configuration editor is opened and an experts settings section, which is hidden beneath an expert settings button in the configuration editor.

Security fix to prevent Cross-Site Request Forgery (CSRF) attacks to succeed.

Creating line charts could lead to a NoSuchMethodError and would prevent generating such reports at all. This could only be fixed by upgrading to a newer charting library version.

Sorting the history view of incidents by clicking on the "value" column is no longer broken.

The format file for the P8 server error logfiles did not parse entries correctly if they ended with a Java exception. This has been fixed.

Some existing Knowledgebase entries did not show up in the details view of an incident although they should. Now this is corrected and these entries are also shown. Mostly the descriptions for P8 server error log entires were affected by this issue.

Lists in the sidebar and portlets scrolled spontaneously to the top if there are automatic refreshes or updates in the UI. This is no longer the case if one has scrolled down.

Datacap Pages Queued Probe can now create more than one sample if needed. So the probe will return samples in such a case instead of returning none and failing silently.

Invalid hostnames in the SNMP configuration no longer generate very technical exceptions, but understandable and helpful error messages.

The probes Icc4SapArchiveStatus and Icc4SapServerStatus failed on UNIX and Linux systems due to an incorrect directory name. This has been fixed.

Datacap probes are now working properly with Oracle.

Icc4SapArchiveStatus probe no longer fails when authentication type is selected or more than one archive is defined in the INI file.

Icc4SapProcess situation no longer fails with the error message The following processes do not have the specified number of instances […​].

The OracleProcesses probe had erroneously returned zero for a successful case instead of one. This is now fixed.

Microsoft Internet Explorer 11 only: Filtering a situation group by a tag resulted in an empty situation list. This is now fixed. The filter works as expected and shows all situations that match the tag used for filtering.

The communication between server and agent is now TLSv1.2 encrypted. Lower encryption protocols are no longer supported. So all agents prior to this fix will not be able to connect to the server after it was updated. All agents must be updated, too, to allow them to reconnect to the server.

It is now possible to have more than one dashboard in the GUI. Dashboards can be created, named, and deleted. Each dashboard can have a different set of portlets.

The welcome page of the product now shows a "Getting Started" section to help a new user to configure a good monitoring of his/her environment.

First probes for monitoring IBM ODM were added to the product. These probes are using the REST API of the RES server.

For MS Windows Event Log monitoring a so called flow limiting or throttling of incidents can now be defined. This is equal to the throttling of incidents for text based logfiles.

The probes Icc4SapArchiveStatus and Icc4SapServerStatus failed on UNIX and Linux systems due to an incorrect directory name. This has been fixed.

Directory traversal vulnerability in ESM 5.5. IF4 does no longer exist.

A pre-installed Java under MS Windows defining JAVA_HOME no longer interferes with the embedded Java used by ESM.

No more connection failures (ERRORCODE=-4228 SQLSTATE=null) with the newest JDBC driver for DB2 11.1.4.4. So using the older driver is no longer needed to connect to a DB2 11.1.4.4.

Optional automatic compression of the embedded database was added to the start-up procedure. For details ask your support personnel.

First probes for monitoring IBM BPM were added to the product. These probes are using the REST API.

The sidebars are now preserving their horizontal size between view changes and sessions.

After deleting an entry in the sidebar the details of the deleted item are also removed from the UI.

Closing of the details aka context view area for incidents is now possible without closing the complete view.

Situation Groups that are defined by more than one tag are now showing all relevant situations again.

New wizard to configure subsystems. The new wizard guides through the process of defining a complete set of related subsystems. For details the updated Configuration Guide.

Cumulative security update of Apache Karaf and bundles Jetty server.

Problems of Datacap probes on Oracle based system fixed.

Icc4SapArchiveStatus probe no longer fails when authentication type is selected or more than one archive is defined in the INI file.

Only relevant KnowledgeBase entries are now shown for cepest based probes with numeric values. Before this fix, all KnowledgeBase entries associated with the subsystem used by the probe were listed in the details view.

Icc4SapProcess situation no longer fails with the error message The following processes do not have the specified number of instances […​].