<PRE>
======================================================================
Halcyon PrimeAlert (R) for Sun Application Server
Version 2.2.0
Release Notes
======================================================================
Copyright (c) 1998-2008 Halcyon Monitoring Solutions, Inc.
Note: For the latest information on this product, visit:
http://www.HalcyonInc.com
Refer to the README file(s) for all other information.
Versions Documented Below: 2.2.0 2.1.0 2.0.0b 2.0.0a 2.0.0 1.0.0b
1.0.0a 1.0.0
======================================================================
VERSION/DATE: 2.2.0/2008-04-04
ENHANCEMENTS
----------------------------------------------------------------------
- Added support for Sun Java System Application Server 9.x.
- Added support for PrimeAlert Neuron.
- Added monitoring of Total Requests for HTTP Server Traffic Summary.
- Added monitoring of Server Instance and Domain JVMs for Application
Server 9.x and above.
BUG FIXES
----------------------------------------------------------------------
- Full PRM support has now been enabled for the module. As a result
of this fix, a number of *.dat files in /var/opt/SUNWsymon/cfg may
be updated (see UPGRADE STRATEGY below). Please note that this
does not require user intervention or input.
- Full support for Sun MC Group Operations has been enabled for the
module.
- Instance sub-module titles and alarm messages are now more
descriptive of what is being monitored and where the alarm is
occurring in the module.
- The EJB tables for the server instances are now supported for PRM
and alarming is fully functional as well.
- Changing the location of the log file for any appserver entity
(domain, instance, node-agent) will now be reflected in the module
and be monitored appropriately.
- The module will now fully support domain, server instance and
node-agent names which include a dot '.'.
UPGRADE STRATEGY
----------------------------------------------------------------------
- If upgrading from 2.0.0 and the combination of
SAS_[dasHost]_[domainName]_[appServerInstanceName] is currently
longer than 36 characters for a given server instance, the new
sub-module instance names will not match the old ones. If this is the
situation, and you wish to upgrade to 2.2.0, you will need to
unload your previous modules and then load instances of the new
module. Note that you will lose previous alarm thresholds and
monitoring information.
- A number of *.dat files in /var/opt/SUNWsymon/cfg may be updated so
that they work properly with PRM. Prior to the update, all the
*.dat files will be copied to backup files in /var/opt/SUNWsymon/cfg.
The backup files will maintain the same name as the original file,
except will have "-[currentDate]-[currentTime].old" appended to them.
Please note that this does not require user intervention or input.
- If you have previously installed patch 3695 for the Application
Server module, you must uninstall it before upgrading/installing
HALSunAppServerJMX version 2.2.0.
KNOWN PROBLEMS
----------------------------------------------------------------------
- The module is unable to handle certain characters, such as space
and "-", in the user name and/or password load parameters.
Although Application Server appears to allow such characters,
they will cause errors if specified as part of one of the load
parameters.
- Loading multiple instances monitoring the same DAS host and domain
may cause LogFileMonitor and ProcessMonitor sub-modules to behave
unpredictably.
- When the module is disabled then enabled, LogFileMonitor and
ProcessMonitor sub-modules may be temporarily disabled.
- If the module is loaded with the incorrect parameters, the file
name for the Server log may be detected incorrectly. Module edits
may fail to update the section with the correct file name.
The work around for this problem is to unload the module and load
it again with the desired parameters.
- The application server log may grow at a fast rate if data
acquisition requests made by the module result in warnings
being written to the server log.
The work around is to monitor the size of the log and rotate it
as needed.
- When the module is loaded, it starts a JMX listener service for
each Server Instance to be monitored.
The DAS is queried to determine the port number used by each
Server Instance. If not found, the Server Instance is considered
to be configured but not running, and a default (dummy) port
number is used as a placeholder for the non-running Server
Instance.
When the Server Instance comes online again, a valid port will
be retrieved and a new listener will be started using that port.
This will halt the previously running listener for that Server
Instance.
Warning messages indicating this process are displayed in the
agent.log and can be safely ignored. Examples of the text
appear below:
Examples of the text appear below:
warning Feb 22 17:27:30 agent
runCommand - shell .services.io.JMXshell.toe5620 has died
warning Feb 22 18:47:47 agent
runCommand - shell .services.io.SAS_terror_domain1_
server1_sh.toe6924 has died
- When the module is loaded (either initially or after the agent
is restarted) a series of warning messages may appear in the
log file for each server instance being monitored. The
text of these messages is:
'refreshValueAndTrap: 9 - invalid command name "refreshValue"'
These warnings are caused by the agent attempting to refresh
server instance entries that have yet to be fully loaded.
This is only a timing issue which does not affect the
functionality of the module and, as such, these messages can
be safely ignored.
NOTES
----------------------------------------------------------------------
- PRM refers to Sun MC's Performance Reporting Manager.
- If, after installing this module, you install an older module that
includes console packages, you may see some warnings or messages
such as "[conflicting pathname not installed]" in the install log.
You should not worry about this. The problem is that some files
that used to be installed to the console layer (in the older
module) are now installed to the server layer (in this module).
Because the installer sees that there is no console layer package
installed for that module, it will try to install one, which may
present these messages. The functionality of that module should
not be impacted at all.
======================================================================
VERSION/DATE: 2.1.0/2007-12-03
ENHANCEMENTS
----------------------------------------------------------------------
- Added support for ssl connections to Sun Application Server.
- Added support for Sun Management Center 4.0.
- Added monitoring for Server Instance specific JVM Memory metrics.
BUG FIXES
----------------------------------------------------------------------
- For some logfile names an error would be generated by the Logfile
sub-module.
- Server instance, domain or host names that were too long when
combined resulted in sub-module instances that may not have been
unique.
UPGRADE STRATEGY
----------------------------------------------------------------------
- This will be compatible with the previous version with one
exception: If the combination of
SAS_[dasHost]_[domainName]_[appServerInstanceName] is currently longer
than 40 characters for a given server instance. In this case, the
new sub-module instance names will not match the old ones. If this is
the situation, and you wish to upgrade to 2.1.0, you will need to
unload your previous modules and then load instances of the new
module. Note that you will lose previous alarm thresholds and
monitoring information.
KNOWN PROBLEMS
----------------------------------------------------------------------
- Loading multiple instances monitoring the same DAS host and domain
may cause LogFileMonitor and ProcessMonitor sub-modules to behave
unpredictably.
- When the module is disabled then enabled, LogFileMonitor and
ProcessMonitor sub-modules may be temporarily disabled.
- If the module is loaded with the incorrect parameters, the file
name for the Server log may be detected incorrectly. Module edits
may fail to update the section with the correct file name.
The work around for this problem is to unload the module and load
it again with the desired parameters.
- The application server log may grow at a fast rate if data
acquisition requests made by the module result in warnings
being written to the server log.
The work around is to monitor the size of the log and rotate it
as needed.
======================================================================
VERSION/DATE: 2.0.0b/2007-07-16
ENHANCEMENTS
----------------------------------------------------------------------
- Added Solaris 10 zone support.
BUG FIXES
----------------------------------------------------------------------
- Certain JDBC connection pool database name might cause the JDBC
table to not populate.
- For certain data, if the number gets too big, it will cause the
data not to populate.
UPGRADE STRATEGY
----------------------------------------------------------------------
- None. This version is compatible with the previous version.
KNOWN PROBLEMS
----------------------------------------------------------------------
- Loading multiple instances monitoring the same DAS host and domain
may cause LogFileMonitor and ProcessMonitor sub-modules to behave
unpredictably.
- When the module is disabled then enabled, LogFileMonitor and
ProcessMonitor sub-modules may be temporarily disabled.
======================================================================
VERSION/DATE: 2.0.0a/2007-06-12
ENHANCEMENTS
----------------------------------------------------------------------
- The PrimeAlert for Sun App Server JMX module has been tuned to limit
memory growth when used with Java 1.5.
BUG FIXES
----------------------------------------------------------------------
- Under certain conditions (dots in the name of an App Server Instance)
App Server instances may fail to appear in the "Server Instances"
section of the module.
- The use of Sun MC Agent Update results in errors
about xml and agent-update package creation or
deployment errors.
- After installation of App Server JMX module on a Sun MC
Server the agent dies when attempting to load the
Solaris Container Manager Module (SCM). The agent.log
contains the following error:
general parsing error file://localhost/scm-container-d.x;flags=ro 9
couldn't load file pkgsrm.so
ld.so.1 esd: fatal: relocation error: file /usr/lib/libpool.so.1:
symbol xmlNewChild: referenced symbol not found
*** aborting execution ***
======================================================================
VERSION/DATE: 2.0.0/2005-05-13
ENHANCEMENTS
----------------------------------------------------------------------
- The module was modified to take advantage of the new JMX server
which is part of Sun Application Server 8.x.
- The module can now automatically detect all server instances within
a domain and monitor all local instances.
- The module now supports the intel Solaris platform.
BUG FIXES
----------------------------------------------------------------------
- None. This new version is released to support the new monitoring
interface of Sun Application Server v8.
UPGRADE STRATEGY
----------------------------------------------------------------------
- This version is not compatible with the previous version. If you
need to monitor Sun Application Server 7.x, you must use
PrimeAlert for Sun Application Server v1. Note that both versions
of the module can be loaded on the same server at the same time.
- If you want to upgrade from PrimeAlert for Sun Application Server v1
to PrimeAlert for Sun Application Server v2, you will need to
unload your previous modules and load instances of the new module.
Note that you will lose previous alarm thresholds and monitoring
information.
======================================================================
VERSION/DATE: 1.0.0b/2004-08-25
ENHANCEMENTS
----------------------------------------------------------------------
- None.
BUG FIXES
----------------------------------------------------------------------
- The Transactions table in the section Java Transaction Services
will update properly if the Transaction Ids are very long.
UPGRADE STRATEGY
----------------------------------------------------------------------
- None. This version is compatible with the previous version.
======================================================================
VERSION/DATE: 1.0.0a/2004-05-28
ENHANCEMENTS
----------------------------------------------------------------------
- None.
BUG FIXES
----------------------------------------------------------------------
- If the module transitions to the down state, all open alarms for the
module will be closed.
- If the module transitions to the down state, duplicate alarms will
not be issued from the Processes and Server Log sections.
- When the module returns from the down state, there will not be false
alarms from the Processes section indicating that the application
server processes are not running.
UPGRADE STRATEGY
----------------------------------------------------------------------
- None. This version is compatible with the previous version.
======================================================================
VERSION/DATE: 1.0.0/2004-04-30
ENHANCEMENTS
----------------------------------------------------------------------
- None. This is the First Customer Shipment (FCS) release.
BUG FIXES
----------------------------------------------------------------------
- None. This is the First Customer Shipment (FCS) release.
KNOWN PROBLEMS
----------------------------------------------------------------------
- If the module is loaded with the incorrect parameters, the file
name for the Server log may be detected incorrectly. Module edits
may fail to update the section with the correct file name.
The work around for this problem is to unload the module and load
it again with the desired parameters.
- The SNMP sub agent may stop responding if an application server is
added, removed or restarted. As a result, some tables in the
module will not update.
This is a known bug in Sun Java Systems Application Server.
Bug Id 4969425.
Please see Troubleshooting for a solution.
- The application server log may grow at a fast rate if data
acquisition requests made by the module result in warnings
being written to the server log.
The work around is to monitor the size of the log and rotate it
as needed.
---//---
</PRE>