Square Bubble for IIB & MQ Release Notes

Version 1.2.6

Release date 11th May 2018.

This is a bug fix release, which has the following fixes:

  1. MQTT client ids are now guaranteed to be unique
  2. Stop process can sometimes get ignored when thread is in an error state, now they stop correctly
  3. HTTP client (used in IIBREST sources) buffer has been increased to 256KB
  4. Elastic visualization IIB Servers by runMode displays up to 99 Server names

Version 1.2.5

Release date 1st February 2018.

This is a minor update & bug fix release, which has the following enhancements:

  1. Start and stop scripts in the tar.gz package now work in git bash and cygwin on Windows.
  2. Splunk connections default to use TLS v1.2. Previously, it was recommended to add -Dhttps.protocols=TLSv1.2 to the sqrbbl-setup file. You will now have to do this if you wich to use earlier versions.
  3. Splunk Dashboard "IIB Flowstats top 10 by Average Elapsed Time" also displays number of calls and max elapsed time per message flow and message flow node.

Along with the following bug fixes:

  1. Unknown PCF parameters returned from later releases (such as CD) of MQ are returned with their numerical value. The MQ V 9.0.0.2 redistributable client is now included in the package.

Version 1.2.4

Release date 24th August 2017.

This is a minor enhancement release for:

  1. IIB Broker timeouts are now reported in the runMode field. This aids with the at a glance view as well as providing an easier method for detecting Broker timeouts.

Version 1.2.3

Release date 9th May 2017.

This is a minor update & bug fix release, which has the following enhancements:

  1. IIB BROKER.Node sub type & MQ QMGR type now have additional fields sourceName, hostName & hostPort added. The values are taken from the configuration source. It has been added to allow dashboards to be built to determine which specific host a clustered node or queue manager is actually running on. This is similar to the Schroedinger's Cat problem, in that a node or queue manager may be described as being both up and down, simultaneously.

Along with the following bug fixes:

  1. Includes for IIBREST were not being OR'd as they should.
  2. Excessive file handles were being generated when writing to an file based output destination using the IBM JRE.

Version 1.2.2

Release date 19th April 2017.

This is a bug fix release, which has the following fixes:

  1. FLOWSTATS now supports non-application message flow deployments. Specifically message flows deployed outside an application or in a static library are now reported. The application name (ApplicationName field) is set to 'undefined' in these cases. Those from a library will also have a LibraryName field.
  2. Timestamp field is NEVER excluded as a result of Include/Exclude processing.

Version 1.2.1

Release date 17th March 2017.

This is a minor update & bug fix release, which has the following enhancements:

  1. MQ Queues Detailed dashboard has been added to provide greater focus on the QSTATS data. It includes a table of queues, which can be filtered and sorted by high queue depth, number of enqueues or dequeues over a configurable time period.
  2. JMX support has been added to increase monitoring of the square bubble process and to allow a better shutdown process, which releases all connections. All provided stop scripts now use this approach. It is worth noting that the IBM JRE does not work using the attach API, so if you wish to use the IBM JRE, we suggest you either kill the square bubble process (as done previously, and done in the stop script where the JMX call fails) or deploy jolokia which provides a REST api that can be used to trigger the shutdown process.
  3. IBM MQ Versions 7 & 7.1 support added.
  4. IBM Websphere Message Broker Versions 7 & 8 support added. This includes flow stats & resources. An additional attribute (brokerVersion) has been added to a MQ source to differentiate between the versions of broker.
  5. Elasticsearch V5.x, including XPack, is now supported.
  6. Elasticsearch REST API is now supported instead of the transport API. For a default installation this means changing the configured port from 9300 to 9200.
  7. The minimum version of Java required is now back to Java 7, as the dependency on the Elasticsearch transport client which forced this up to Java 8, has now been removed.
  8. The -props option has been added to remove the need to include passwords in the command line. Instead they are loaded from a file.
  9. Daily rolling log config used for .tar.gz/.zip file distributions.
  10. Use of additional jar files is now supported, this is required for security exits or additional monitoring support (such as the use of jolokia).

Along with the following bug fixes:

  1. Improved handling of MQ connection failures. In some failure situations this could result in a tight loop. Now the process sleeps for 5 seconds when a connection is lost before retrying to connect.
  2. Where connections to an IIBREST interface are totally lost, the URL cache is restarted. In normal running all status URLs are requested at the same time, once there is a successful list built. In cases where the interface is not available, this can be excessive and can lead to an OutOfMemoryException being raised.
  3. Splunk MQ At a glance dashboard was not displaying all queues by queue manager. In some cases queues that appear on more than one queue manager were not being displayed.

Version 1.2.0

This is a minor update release, which has the following changes:

  1. The splunk output is contained within a single index, called sqrbbl. You will see lots of sourcetypes within the index.
  2. Support has been added for the Splunk HTTP Event Collector. The output type SPLUNKHEC has been added along with the token attribute.
  3. Username and password is now supported for MQ connections. The name attribute for a Source of type MQ must now match the Queue Manager name in order to connect. Getting this wrong will now result in the message MQJE001: Completion Code '2', Reason '2058' appearing in the sqrbbl.log.
  4. Security Exit class name can also be added to an MQ Source using the securityexit attribute
  5. Detailed MQ tracing is available (using the command line -mqtrace option and editing the logback.xml). This is for detailed analysis of possible issues.

Version 1.1.1

This is the first bug fix release, which has the following fixes:

  1. Q_NAME filters which only appeared in the Excludes section were being ignored. This only applied to the QUEUES and QSTATS data types.
  2. Elastic was complaining about the index template being corrupt. The templates are recreated if this occurs, which allows data to be ingested by elasticsearch V5.

The following features have been changed:

  1. Support for Java 7 has been removed. You need to use Java 8 or above.

Version 1.1.0

This is the first minor update, which added the following features:

  1. Support added for QSTATS from MQ. This can be added to an existing source of type MQ. It acquires its data using the MQ RESET QUEUE STATISTICS command. This should only be done by a single agent, otherwise data will be incomplete, therefore you should decide whether you wish to do this through Square Bubble. To switch off, just add QSTATS to the exclude section of the MQ source.
  2. QSTATS dashboard has been included for both splunk and elastic

The following features have been changed:

  1. Filters in the include section are treated with logical OR operations, previously they were subject to a logical AND (filters in the exclude remain unchanged i.e. subject to a NOT AND or NAND operation). This has been done to enable the output from MQ commands to be reduced to a smaller result set, thus reducing the load on MQ. This is particularly relevant to the QSTATS. Previously, if you wished to report on a subset of queues based on name, you would either have a single include or a set of excludes. Whilst this will achieve the correct output from square bubble, MQ would build output for all queues and throw away those that do not match the filter. This change means you can specify multiple name filters which will make MQ only produce the subset of data required.
  2. The Splunk app must be redeployed on all splunk servers in the chain (whether forwarders or search heads), as the index names have changed. They contain the major/minor versions in the name. Be aware that this may cause the existing data to be lost. Taking a backup is therefore recommended to prevent this.