The
This section summarizes the new features in this release.
The
The
Support for externally controlled VIP. (PFM-7287)
An Engage cluster can now be accessed through an externally managed virtual IP or DNS hostname. Large customers with fixed network requirements can now use a multi-datacenter deployment and not need to support layer-2 adjacency (sub-net spanning) across data centers.
This section provides information about fixes and improvements made to the
The following issues are fixed in Engage Platform 5.5 HF1 (5.5.3-3):
The platform migrator script called in an upgrade to 5.5 no longer clears the yum cache which could lead to failed upgrades on network disconnects; the yum clean command has been removed from the migration script. (PFM-9166)
The following issues are fixed in Engage Platform 5.5 (5.5.3):
Some hostname values will break clustering. (PFM-8304)
Resolution: Prevent the user from entering "mqhost" or "eihost" in the hostname field. The Engage Platform Gold image now uses "engage" as the default hostname.
The issue no longer occurs where a slave node could generate Audit 416 for the Core Queues surpassing the threshold. (PFM-6182)
The issue no longer occurs where the SMTP settings in the audit log were not being synced from the master node to the slave nodes. (PFM-8115)
The issue no longer occurs where case sensitivity in equality methods for objects in the Device Management Library triggered unnecessary updates. (PFM-7211)
This section provides information about known product defects and limitations in the current release.
A cluster of three nodes is not supported. In a cluster of three or more nodes, changing the configuration of the cluster by adding or removing a node causes the master node to failover. (PFM-9112, PFM-9023)
Upon failover, messages in-flight are unable to be delivered. (PFM-9082)
Engage and Voice Server do not currently have the same username management requirements. Currently, the Voice Server supports usernames that are up to 50 characters in length, while Engage supports usernames up to 40 characters in length. The Voice Server supports usernames that are as short as a single character, while Engage requires a username to be a minimum of 3 characters. When a username is created in Engage that does not match the Voice Server requirements, the inconsistency will not be logged and the user will not link to a device. (PFM-8916)
The Active checkbox setting in the Solution Rules is not respected when the "Don't send back to originating adapter" box is checked. (PFM-8846)
Script Execution Engine does not ensure timers are successfully linked during upgrade from Platform v5.3 to a later version. (PFM-8445)
First run the following commands to determine which timers were not properly linked:
sudo systemctl list-timers --all
Any suspect timers should be investigated with:
sudo systemctl status <service>.timer
If the timer has not run or is not active, it has probably not been linked properly.
Workaround is to run the following commands. (Example is for vacuumdb; replace vacuumdb with service name of the problem service.)
# Link vacuumdb timer/service /usr/bin/systemctl link /opt/EXTENSION/conf/.systemd/vacuumdb.service /usr/bin/systemctl link /opt/EXTENSION/conf/.systemd/vacuumdb.timer # Ensure that the vacuumdb timer is wanted by ha_master and standalone /usr/bin/systemctl add-wants ha_master vacuumdb.timer 2>/dev/null /usr/bin/systemctl add-wants ha_standalone vacuumdb.timer 2>/dev/null # Add the enabled file to enable the vacuumdb script. su extension -c "touch /opt/EXTENSION/conf/.systemd/vacuumdb.enabled" # Reload the daemon sudo /usr/bin/systemctl daemon-reload # Start the vacuumdb timer sudo systemctl start vacuumdb.timer
Run the following command to initiate a vacuum analyze immediately:
sudo systemctl start vacuumdb.service
Migration from Platform 5.2.17 to Platform 5.5.0 can leave extension-postgres in a bad state. (PFM-8099)
For the situation where extension-postgres is NOT being installed before the migration is initiated, use this workaround to access the branch that is missed on installation. After the migration has concluded, run the following command:
sudo yum reinstall extension-postgres
When the XMPP interface is restarted, all undelivered messages currently in offline storage are lost. This affects users that Engage Platform identifies as disconnected at the time of a failover. (PFM-283)
When a message is delivered to a user, it is written to the Messages dataset. When a message is sent to a user who is not connected, the message is kept in memory by the interface. If the interface is restarted, all messages which have not been delivered are lost.
Exporting the solution package fails on certain configurations. (PFM-6205)
Package export can fail with a Page Not Found error due to a rule involving the MassAlerts dataset.Attribute transformation RegEx cannot distinguish the appendage of a string literal of a number right after the RegEx attribute. (PFM-7216)
No known workaround.
Audit Log reports an exception if the appliance time is set for far future (i.e., + 10 years) and then reverted back to the current time. (PFM-6250)
Workaround: Remove and recreate the current audit database schema by executing the following commands:
sudo -u extension psql -p 5433 -c "DROP SCHEMA public cascade;" audit sudo -u extension psql -p 5433 -c "CREATE SCHEMA public;" audit sudo -u extension psql -p 5433 -c "CREATE extension hstore;" audit
Redeploy the ei-audit.jar (it may take several minutes until the audit database is restored to a functional state):
sudo su extension; touch /opt/glassfish/glassfish/domains/extension/autodeploy/bundles/ei-audit.jar
The XMPP Certificate Manager is empty. (PFM-6127)
Workaround: Make sure the adapter is running, refresh the adapter page, and wait for the XMPP Certificate Manager to display.
When adding new trusted certificates for adapters, GlassFish must be restarted. (PFM-5949)
Workaround: Obtain certificate chain, upload via Security > Certificates, and restart GlassFish.
Removing Mirth also removes extension-install and extension-support-keys packages. (PFM-4499)
Workaround: Do not remove Mirth if not using it; disable the service. If Mirth is removed, reinstall the extension-support-keys and extension-install packages.
Datasets created from new adapters fail to save and deactivate the "Save" button. (PFM-1942)
Workaround: Create the required datasets before configuring the adapters.
GlassFish locks up during Felix deployment of JTAPI libraries. (PFM-1882)
Workaround: Avoid unnecessary configuration changes to the adapters in production deployments. If a configuration change is necessary, schedule a maintenance window with the customer anticipating that a GlassFish restart may be required to return the server to an operating state.
Clinical alarms can be delayed depending on how recipients are computed. (PFM-985)
Workaround: Add the following filters followed by a vacuum analyze:
Handler queues do not start up within 5 minutes of rebooting appliance. (PFM-459)
Workaround: Wait for Adapter Services to display the correct status.
SNMP Community name does not support special characters, "$",“%”, “/”, “\”: (PFM-303)
Workaround: First enter the community string without the special character from the SNMP tab.
Login into system vi CLI, then:
cd /etc/snmp cp snmpd.conf snmpd.conf.orig vi snmpd.conf
Edit the configuration adding the special characters, e.g.:
rocommunity Solarwind$ default -V eiView trapcommunity Solarwind$ trap2sink 10.159.77.225 Solarwind$ 162
Restart the SNMP daemon:
sudo /etc/init.d/snmpd restart
Large workflow queries can cause OutOfMemory errors, and GlassFish to become unresponsive. (PFM-273)
Workaround: Follow proper change control methods when configuring new workflows.
If a CSV file contains an "id" column, the sequence for the table will be left out of sync, resulting in an error in Audit Log with stack trace containing "ERROR duplicate value violates unique constraint" after completing a CSV import. (PFM-266)
Workaround: If the CSV file contained an "id" column, it will cause the table serial sequence to get out of sync until the sequence catches up to the last CSV ID imported. Do not include the "id" column when importing CSV files.
During import of ResponderSync adapter, clicking the button allows adapter creation without mandatory fields. The error message "Staff Assignment Service URL undefined already used in another configuration." is displayed if the ResponderSync adapter is not configured correctly. (PFM-263)
Workaround: The ideal workaround is to re-image the appliance or VM to begin a fresh import action, and avoiding the issue during the configuration of the ResponderSync adapter.
Alternate: It is possible to salvage the current work from the import using the following steps:
Cloning rules may take a significant amount of time with no visual indicator to the user. This may lead the user to attempt the clone several times and cause unexpected results, i.e., clones of clones. (PFM-242)
Workaround: After clicking the button to clone the rules, wait until the clone finishes completely before clicking the button again.