Contents
Introduction
This document contains Voice Portal 5.1 Service Pack 2 product information that is not included in the product documentation. This document highlights known issues about the Voice Portal product along with workarounds that are available.
Note: Before continuing, check the Avaya support Web site at http://support.avaya.com for an updated version of this document.
Problems Fixed
Number | Problem Description |
---|---|
wi00955165 | Preserve files for managed applications during Avaya Linux upgrade |
wi00954417 | Script vpupgrade.sh does not restore folder avpom during Avaya Linux upgrade |
wi00953056 | Need to update httpd package on Avaya Linux |
wi00953053 | Need to update kernel package on Avaya Linux |
wi00949605 | After DTMF input, background noise sometimes causes false barge-in of next prompt |
wi00945208 | Need to update rpm packages |
wi00944629 | Update Avaya Linux with latest tzdata RPM |
wi00942355 | Create new Avaya Linux based on RHEL 5.7 |
wi00940950 | Modify default.ccxml script to handle Special Information Tone (SIT) correctly |
wi00937798 | VoiceXML 2.0 assertion #362 fails when n-best recognition is disabled |
wi00937525 | Problem executing next nodes if consultative transfer fails due to line busy |
wi00937329 | Update the copyright date on the legal notice |
wi00936632 | Core dump when using MRCPv2 |
wi00936550 | Need to update Apache httpd packages |
wi00934295 | Tighten Apache Axis2 security |
wi00934293 | Need to update sysstat package |
wi00934211 | MediaManager core dump |
wi00933985 | Upgrade to Postgres 8.3.14 |
wi00933892 | Fix several issues in script getmpplogs.sh and related packcore.sh |
wi00930673 | Need to update libXfont package |
wi00930672 | Need to update libpng in Avaya Linux |
wi00930669 | Need to update bash in Avaya Linux |
wi00928411 | Slow H.323 station removal due to application server polling |
wi00927000 | H.323 calls fail when switch is configured to require media encryption |
wi00925409 | VoiceXML Interpreter core dump while processing recognition result |
wi00925404 | Update client truststore with latest WebLM security certificate |
wi00920612 | SpiderMonkey core dump |
wi00908766 | Need to update curl |
wi00908761 | Need to update Java |
wi00905502 | Need to update gd package |
wi00905500 | Need to update krb5-libs |
wi00898402 | SessionManager core dump with MRCPv2 |
wi00896936 | Update Install Agent to handle new certificate |
wi00893054 | Core dump on completed transfer when SIP connection is bad |
wi00891325 | SIP ALLOW header erroneously indicates support for UPDATE method |
wi00890030 | Log highest confidence level and matched utterance when n-best is enabled |
wi00886417 | Need to update python packages |
wi00886405 | VoiceXML Interpreter should try to reconnect if http socket connection is closed by app server in the middle of a call |
wi00883414 | H.323 shuffling sometimes results in a core dump |
wi00881061 | Applications running on WebLogic 11g sometimes fail to retrieve configurable application variables |
wi00879454 | Maximum length and maximum silence parameters not handled correctly when recording |
wi00869847 | Need to update openldap packages in Avaya Linux |
wi00869797 | Need to update dbus packages in Avaya Linux |
wi00869400 | Need to update samba client in Avaya Linux |
wi00866322 | Inbound fax detection sometimes fails after a re-INVITE |
wi00866302 | Incorrect handling of SIP timers causes ICR to re-queue calls |
wi00860179 | Problem with call classification timing |
wi00846396 | Core dump when recording is terminated because of maximum length |
wi00840977 | VoiceXML Interpreter should always include "filename" parameter when generating a "Content-Disposition" header |
wi00840857 | Voice Portal fails to send SDP answer when rejecting an SRTP medium |
wi00839022 | Need to update Linux kernel |
wi00837740 | Add capability to specify a pre-buffer time for recording |
wi00837533 | DTMF inter-digit timeout not handled correctly if type ahead is enabled |
wi00833233 | Modifying MPP MRCP network address causes core dump |
wi00498327 | Enable application to send a re-INVITE to promote a no media call |
Product Information
The Voice Portal Documentation Library is available on the Voice Portal service pack DVD. To access the Voice Portal Documentation Library, place the service pack DVD into a Windows system and use Internet Explorer to view the file \Documentation\VoicePortalDocLibrary\_StartHere.html. From this page, you can access the following information:
Printable PDF files for most of the above are accessible from the page titled About the Avaya Voice Portal library and from the Print guides section of the table of contents.
Also available from the Print guides section of the table of contents are the white papers listed below:
Note: License agreements for each of the third-party components that ship with Voice Portal can be found in the Voice Portal DVD directory /Licenses.
Installation Notes
Important: Before installing or upgrading Voice Portal, please review the Known Issues section in this document for issues that are not addressed in the product documentation. Several of the issues listed there are related to installation.
Important: Voice Portal 5.1 SP2 requires Red Hat Enterprise Linux Server release 5.6 or later, or Avaya Enterprise Linux version RHE5.7-AV16.0VP5 or later.
This service pack can be used to do a fresh install of a new Voice Portal system. You do not need to install Voice Portal 5.1 before installing this service pack.
For detailed installation procedures, see either Implementing Voice Portal on multiple servers or Implementing Voice Portal on a single server, whichever is appropriate for your installation. These guides also contain important installation worksheets that should be filled out before starting the Voice Portal installation.
Important: Before starting an upgrade, you should back up your Voice Portal database. If the upgrade fails for any reason, you will need this backup to restore your system to its prior state.
For detailed upgrade procedures see Upgrading from Voice Portal 4.1 or 5.0 to Release 5.1. The procedure to upgrade a system to this service pack is the same as the procedure to upgrade a system to Voice Portal 5.1. Note that because this service pack requires Red Hat Enterprise Linux version 5.6 or later, or Avaya Enterprise Linux version RHE5.7-AV16.0VP5 or later, you will likely need to upgrade the Linux operating system before upgrading Voice Portal.
Important: Upgrading from Voice Portal 4.1 or 5.0 to Release 5.1 describes multiple methods for installing the Voice Portal software. If possible, you should use the "Automatic upgrade" method that has you run a script called autoupgradevp. This will avoid a known issue that can cause all of your MPPs to go into the Not Responding state if you upgrade your Primary VPMS using the script installvp.
Important: You cannot use the Software Upgrade web page of the VPMS to upgrade to this service pack.
If you are upgrading a Voice Portal system that was originally installed with Voice Portal 4.1 SP3 or earlier, then you need to upgrade the schema of your Voice Portal internal database. This schema upgrade is performed on the Primary VPMS after you upgrade to Voice Portal 5.1.
Note: If you are upgrading a system that was originally installed with Voice Portal 4.1 SP4 or later, then you do not need to perform the procedure below. However, if the system was originally installed with an earlier release and then upgraded to 4.1 SP4 or later, then you do need to perform the procedure below.
To upgrade the schema of your internal database:
/sbin/service vpms stop |
sudo -u postgres psql -d VoicePortal -c "alter table sdproperty alter column value type character varying(1024)" |
/sbin/service vpms start |
Known Issues
The Voice Portal installation program must be run from an account that has root privileges. If you are installing or upgrading Voice Portal on a server that uses the Avaya provided version of Linux and you do not have root access, you should open a request with Avaya Support.
On a Primary VPMS server, the Voice Portal database might grow to be quite large. The report data generated automatically by Voice Portal causes the database to grow by approximately 4GB for every one million calls processed. If your applications use the Voice Portal application logging feature, the database grows even faster.
By default, the Voice Portal database resides in the directory /var/lib/pgsql/data. If you are a software-only customer, when you install Red Hat Enterprise Linux on your VPMS server you should ensure that the disk partitions you define provide adequate space for the Voice Portal database. Note that the default partition scheme provided by Red Hat Enterprise Linux has proven adequate for most Voice Portal installations. It results in the Voice Portal database residing in a partition that spans most of the hard disk.
On an MPP server, the MPP logs directory might grow to be quite large. This is particularly true if your applications use the VoiceXML <log> tag or CCXML <log> tag to generate application log data.
By default, the MPP logs reside in the directory /opt/Avaya/VoicePortal/MPP/logs. If you are a software-only customer, when you install Red Hat Enterprise Linux on your MPP server you should ensure that the disk partitions you define provide adequate space for the MPP logs. Note that the default partition scheme provided by Red Hat Enterprise Linux has proven adequate for most Voice Portal installations. It results in the MPP logs being in a partition that spans most of the hard disk.
Be aware that the partition scheme provided by Avaya Enterprise Linux results in the MPP logs residing in a relatively small disk partition. If you encounter space problems, see the help topic Administering Voice Portal > Media Processing Platforms > Moving the MPP logs to a different location for instructions on how to change where MPP logs are stored.
In some instances, when Red Hat Enterprise Linux Server 5.x is installed on a server where the console monitor is connected through a KVM switch the Red Hat install program does not correctly identify the type of monitor connected. One consequence of this is that the Red Hat install program runs in console mode, rather than the typical graphical mode. Another consequence is that after Red Hat Enterprise Linux Server 5.x is installed, it runs in console mode rather than graphical mode.
To configure Red Hat Enterprise Linux Server 5.x to run in graphical mode:
system-config-display |
id:5:initdefault: |
In some instances, when a Voice Portal 4.x system installed on an installation of Red Hat Enterprise Linux ES4 that does not include a graphical user interface is upgraded to Red Hat Enterprise Linux 5.x, the Linux console gets flooded with errors. This makes it difficult to log into the system.
Note: This problem can occur with both Red Hat Enterprise Linux and the version of Linux provided by Avaya.
To avoid this problem, follow the procedure below before upgrading to Red Hat Enterprise Linux 5.x:
Note: If you do not follow the procedure above before upgrading to Red Hat Enterprise Linux 5.x and you encounter the problem of the console being flooded with errors, you can resolve the issue by logging into the system remotely using an SSH client and then performing the procedure above. Alternatively, you can log into the system remotely and install Voice Portal.
If you are upgrading a system that uses the Avaya provided version of Linux, when you run the Linux upgrade script vpupgrade.sh you may see a message similar to the following:
cpartition: ERROR - boot label Uya1 from "lilo -v -q" output does not equate to / or /root2 in "mount" output
You may safely ignore this message.
If you are upgrading an MPP server that uses the Avaya provided version of Linux, when you run the Linux upgrade script vpupgrade.sh you may see a message similar to the following:
Error: Detected the MPP process, SessionManager, running.
Please use the VPMS to stop this MPP and change its
'Operation Mode' to 'Offline'.
You must take the MPP out of service in order to prevent vpupgrade.sh from aborting. As described in the error message, the best way to take an MPP out of service is to go to the MPP Manager page of the VPMS, stop the MPP, and then set the MPP mode to Offline. However, if the VPMS is reporting the state of the MPP as Not Responding, then that procedure will not work.
If you cannot stop the MPP through the VPMS web interface, then follow the procedure below:
/sbin/service mpp stop |
During the portion of the install program where prerequisites are installed, a message similar to the following might be displayed:
| | >>>>>>Starting RPM installation: 'rpm -U --replacepkgs *.rpm'
| | warning: httpd-2.2.3-53.el5_7.1.i386.rpm: Header V3 DSA signature:
| | NOKEY, key ID 37017186
| | error: Failed dependencies:
| | httpd = 2.2.3-53.el5 is needed by (installed) httpd-manual-2.2.3-53.el5.i386.
| | >>>>>>RPM Installation failed: Exit Code: 2
| | ====================================
| | RPM installation check: Expecting 'Found' >= 'Expected'.
| | Expected: httpd-2.2.3-53.el5_7.1.i386.rpm
| | mod_ssl-2.2.3-53.el5_7.1.i386.rpm
| | openssl-0.9.8e-20.el5.i386.rpm
| | Found: httpd-2.2.3-53.el5 Out of Date
| | mod_ssl-2.2.3-53.el5 Out of Date
| | openssl-0.9.8e-20.el5 Passed
If prerequisite installation fails, the installation will not continue.
To resolve this issue:
rpm -e --nodeps httpd-manual |
During the portion of the install program where prerequisites are installed, a message similar to the following might be displayed:
| | >>>>>>Starting RPM installation: 'rpm -U --replacepkgs *.rpm'
| | warning: php-4.3.9-3.22.9.i386.rpm: V3 DSA signature: NOKEY, key ID
| | db42a60e
| | error: Failed dependencies:
| | php = 4.3.9-3.22.4 is needed by (installed) php-ldap-4.3.9-3.22.4.i386.
| | >>>>>>RPM Installation failed: Exit Code: 3
| | ====================================
| | RPM installation check: Expecting 'Found' >= 'Expected'.
| | Expected: php-pear-4.3.9-3.22.9.i386.rpm
| | php-4.3.9-3.22.9.i386.rpm
| | php-domxml-4.3.9-3.22.9.i386.rpm
| | Found: php-pear-4.3.9-3.22.4 Out of Date
| | php-4.3.9-3.22.4 Out of Date
| | php-domxml-4.3.9-3.22.9.i386.rpm Not Installed
If prerequisite installation fails, the installation will not continue.
To resolve this issue:
rpm -e --nodeps php-ldap |
During the portion of the install program where Voice Portal is actually installed and the install program displays Installation Progress, a message similar to the following might be displayed:
avc: granted {setbool} for pid=xxxx comm='setsebool' scontext...
This message does not indicate that an error has occurred. You may safely ignore this message.
After you upgrade your Primary VPMS, all of your MPPs might go into the Restart Needed state. If this happens, the MPPs will continue to process calls normally but will need to be restarted at some point. You can either restart the MPPs immediately after you upgrade the Primary VPMS, or you can wait and restart each MPP after it is upgraded.
After you upgrade your Primary VPMS your Auxiliary VPMS will no longer be able to communicate with it. To resolve this issue, upgrade your Auxiliary VPMS to the same release of Voice Portal that your Primary VPMS is running.
Note: While the Auxiliary VPMS is not able to communicate with the Primary VPMS, it will still be able to communicate with the MPP servers. Therefore, outcall requests made through the Application Interface web service on the Auxiliary VPMS will continue to succeed.
After you upgrade an Auxiliary VPMS (formerly Secondary VPMS) server to Voice Portal 5.1, you must manually restart the Apache httpd server.
To restart the httpd server:
/sbin/service httpd restart |
If you upgrade a system that includes a Secondary VPMS to Voice Portal 5.1, on the VPMS Servers page of the VPMS web application the Secondary VPMS will be renamed VPMS2. This behavior is by design. In Voice Portal 5.1, the term Secondary VPMS has been replaced by Auxiliary VPMS.
After you upgrade Avaya Linux on a single server system that includes a co-resident application server, the co-resident application server should start automatically but it does not.
To resolve this issue:
/sbin/service appserver start |
When you install either a Voice Portal Primary VPMS or a Voice Portal Auxiliary VPMS, the system file /etc/mailcap is corrupted. This corruption is caused by the Java Development Kit that is installed by Voice Portal.
Note: The file /etc/mailcap is used by Linux to determine which program should open a particular file based on the file's MIME type.
To resolve this issue:
text/html; /usr/bin/htmlview %s ; copiousoutput\napplication/x-java-jnlp-file; /usr/bin/javaws %s |
text/html; /usr/bin/htmlview %s ; copiousoutput |
application/x-java-jnlp-file; /usr/bin/javaws %s |
If you are upgrading a Voice Portal system that runs on Oracle Linux, then you must follow the procedures below before upgrading Voice Portal. If you do not follow these procedures, the prerequisite installer portion of the Voice Portal install will fail due to some RPM naming inconsistencies between Oracle Linux and Red Hat Enterprise Linux.
Note: This issue with Oracle Linux only affects upgrades. You may do a fresh install of this service pack on Oracle Linux without performing the extra procedures below.
From any machine with Internet access:
httpd-2.2.3-53.0.2.el5_7.3.i386.rpm |
mod_ssl-2.2.3-53.0.2.el5_7.3.i386.rpm |
xorg-x11-server-Xvfb-1.1.1-48.76.0.1.el5_7.5.i386 |
On each Voice Portal single-server system (VPMS+MPP) to be upgraded:
rpm -U httpd-2.2.3-53.0.2.el5_7.3.i386.rpm mod_ssl-2.2.3-53.0.2.el5_7.3.i386.rpm xorg-x11-server-Xvfb-1.1.1-48.76.0.1.el5_7.5.i386.rpm |
cp -rp /media/cdrom /tmp/voiceportalmedia |
cd /tmp/voiceportalmedia/external/helper |
<rpm> chkfontpath-1.10.1-1.1.i386.rpm;libXau-1.0.1-3.1.i386.rpm;libX11-1.0.3-11.el5.i386.rpm;libXdmcp-1.0.1-2.1.i386.rpm;xorg-x11-filesystem-7.1-2.fc6.noarch.rpm;libfontenc-1.0.2-2.2.el5.i386.rpm;libXfont-1.2.2-1.0.4.el5_7.i386.rpm;libFS-1.0.0-3.1.i386.rpm;xorg-x11-xfs-1.0.2-4.i386.rpm;xorg-x11-font-utils-7.1-2.i386.rpm;xorg-x11-server-Xvfb-1.1.1-48.76.el5_7.5.i386.rpm;xorg-x11-fonts-base-7.1-2.1.el5.noarch.rpm </rpm> |
<rpm> chkfontpath-1.10.1-1.1.i386.rpm;libXau-1.0.1-3.1.i386.rpm;libX11-1.0.3-11.el5.i386.rpm;libXdmcp-1.0.1-2.1.i386.rpm;xorg-x11-filesystem-7.1-2.fc6.noarch.rpm;libfontenc-1.0.2-2.2.el5.i386.rpm;libXfont-1.2.2-1.0.4.el5_7.i386.rpm;libFS-1.0.0-3.1.i386.rpm;xorg-x11-xfs-1.0.2-4.i386.rpm;xorg-x11-font-utils-7.1-2.i386.rpm;xorg-x11-fonts-base-7.1-2.1.el5.noarch.rpm </rpm> |
Important: When deleting xorg-x11-server-Xvfb-1.1.1-48.76.el5_7.5.i386.rpm; be sure to delete the semicolon at the end.
<rpm> httpd-2.2.3-53.el5_7.1.i386.rpm;mod_ssl-2.2.3-53.el5_7.1.i386.rpm;openssl-0.9.8e-20.el5.i386.rpm </rpm> |
<rpm> openssl-0.9.8e-20.el5.i386.rpm </rpm> |
Important: When deleting httpd-2.2.3-53.el5_7.1.i386.rpm;mod_ssl-2.2.3-53.el5_7.1.i386.rpm; be sure to delete the semicolon at the end.
cd /tmp/voiceportalmedia |
bash autoupgradevp |
On each VPMS to be upgraded:
rpm -U httpd-2.2.3-53.0.2.el5_7.3.i386.rpm mod_ssl-2.2.3-53.0.2.el5_7.3.i386.rpm xorg-x11-server-Xvfb-1.1.1-48.76.0.1.el5_7.5.i386.rpm |
cp -rp /media/cdrom /tmp/voiceportalmedia |
cd /tmp/voiceportalmedia/external/helper |
<rpm> chkfontpath-1.10.1-1.1.i386.rpm;libXau-1.0.1-3.1.i386.rpm;libX11-1.0.3-11.el5.i386.rpm;libXdmcp-1.0.1-2.1.i386.rpm;xorg-x11-filesystem-7.1-2.fc6.noarch.rpm;libfontenc-1.0.2-2.2.el5.i386.rpm;libXfont-1.2.2-1.0.4.el5_7.i386.rpm;libFS-1.0.0-3.1.i386.rpm;xorg-x11-xfs-1.0.2-4.i386.rpm;xorg-x11-font-utils-7.1-2.i386.rpm;xorg-x11-server-Xvfb-1.1.1-48.76.el5_7.5.i386.rpm;xorg-x11-fonts-base-7.1-2.1.el5.noarch.rpm </rpm> |
<rpm> chkfontpath-1.10.1-1.1.i386.rpm;libXau-1.0.1-3.1.i386.rpm;libX11-1.0.3-11.el5.i386.rpm;libXdmcp-1.0.1-2.1.i386.rpm;xorg-x11-filesystem-7.1-2.fc6.noarch.rpm;libfontenc-1.0.2-2.2.el5.i386.rpm;libXfont-1.2.2-1.0.4.el5_7.i386.rpm;libFS-1.0.0-3.1.i386.rpm;xorg-x11-xfs-1.0.2-4.i386.rpm;xorg-x11-font-utils-7.1-2.i386.rpm;xorg-x11-fonts-base-7.1-2.1.el5.noarch.rpm </rpm> |
Important: When deleting xorg-x11-server-Xvfb-1.1.1-48.76.el5_7.5.i386.rpm; be sure to delete the semicolon at the end.
<rpm> httpd-2.2.3-53.el5_7.1.i386.rpm;mod_ssl-2.2.3-53.el5_7.1.i386.rpm;openssl-0.9.8e-20.el5.i386.rpm </rpm> |
<rpm> openssl-0.9.8e-20.el5.i386.rpm </rpm> |
Important: When deleting httpd-2.2.3-53.el5_7.1.i386.rpm;mod_ssl-2.2.3-53.el5_7.1.i386.rpm; be sure to delete the semicolon at the end.
cd /tmp/voiceportalmedia |
bash autoupgradevp |
On each MPP to be upgraded:
rpm -U httpd-2.2.3-53.0.2.el5_7.3.i386.rpm mod_ssl-2.2.3-53.0.2.el5_7.3.i386.rpm |
cp -rp /media/cdrom /tmp/voiceportalmedia |
cd /tmp/voiceportalmedia/external/helper |
<rpm> httpd-2.2.3-53.el5_7.1.i386.rpm;mod_ssl-2.2.3-53.el5_7.1.i386.rpm;openssl-0.9.8e-20.el5.i386.rpm </rpm> |
<rpm> openssl-0.9.8e-20.el5.i386.rpm </rpm> |
Important: When deleting httpd-2.2.3-53.el5_7.1.i386.rpm;mod_ssl-2.2.3-53.el5_7.1.i386.rpm; be sure to delete the semicolon at the end.
cd /tmp/voiceportalmedia |
bash autoupgradevp |
In Voice Portal 5.1 the SIP domain name received from the SIP proxy on an incoming call is validated against the value of SIP Domain entered on the Add SIP Connection (or Change SIP Connection) page of the VPMS. If the domain names do not match, Voice Portal will not answer the call. Previous versions of Voice Portal did not validate the SIP domain name received from the SIP proxy.
Because of this change, it is possible that an improperly configured system that worked with previous releases of Voice Portal stops working when upgraded to Voice Portal 5.1.
When upgrading a system to Voice Portal 5.1, verify that the SIP domain name configured on Voice Portal matches the SIP domain name configured on any attached SIP proxies.
An enhancement to the call classification code in Voice Portal 5.1 means that CCXML applications that make outcalls may now receive the event <start_of_voice>. This event occurs after an outcall is answered, at the point where Voice Portal starts trying to determine whether the call was answered by a human or an answering machine. Previous versions of Voice Portal did not send the event <start_of_voice>.
Because of this change, CCXML applications that worked with previous releases of Voice Portal may stop working when the system is upgraded to Voice Portal 5.1. Applications that were modeled after the Voice Portal page default.ccxml from previous releases are likely to suffer from this problem. These applications need to be modified to properly handle the event <start_of_voice>.
Note: In most cases applications can simply ignore the event <start_of_voice>. Where problems occur is when an application looks for a certain set of events and aborts if it receives an unexpected event. These applications need to recognize <start_of_voice> as an expected event.
The version of the sample application BloominBloomsLite that shipped with Dialog Designer 5.0 does not work correctly with Voice Portal 5.1. If you use BloominBloomsLite, use the version that ships with Dialog Designer 5.0 SP1, or newer.
Video applications written for Voice Portal 5.0 might not operate correctly on Voice Portal 5.0 SP1 and later. In Voice Portal 5.0, if a video is played in parallel with audio that is longer than the video, then at the end of the video the last frame will remain on the screen while the audio finishes. Starting with Voice Portal 5.0 SP1, the screen will go blank when the video finishes playing, unless the application writer specifies a video duration that is longer than the actual length of the video.
Note: Application writers should refer to the version of the sample application BloominBloomsLite that ships with Dialog Designer 5.0 SP1, or newer, for examples of video application best practices.
If you have Proactive Outreach Manager (POM) installed on your Voice Portal system and you upgrade Voice Portal, POM may stop working. See the section Upgrading Voice Portal and/or Avaytised linux in the document Implementing Proactive Outreach Manager for instructions on how to make POM work again.
Note: Implementing Proactive Outreach Manager can be found in the Print Guides section of the POM Help system.
When you install the Primary VPMS, a root security certificate is generated. Voice Portal 5.1 generates a root certificate that is valid for ten years. However, some previous versions of Voice Portal generated a root certificate that was valid for only one year. If your root certificate is about to expire, you may see a message similar to the following on the VPMS home page:
SIP: The root CA certificate will expire in 30 days.
To generate a new root certificate:
cd $AVAYA_HOME/Support/VP-Tools |
bash UpdateRootCertificate |
Important: After generating a new root certificate, you must restart all of your MPPs. Also, if you have installed the Voice Portal root certificate on any other servers, such as SIP proxies, in order to support TLS connections, then you must update all of those other servers with the new Voice Portal certificate.
People calling into applications running on a very heavily loaded MPP may occasionally experience delays of several seconds between prompts. This symptom can be caused by file system latencies within Linux. To address this issue, try one or more of the solutions below:
Or
Or
Or
If your application refers to a media a file (for example, a .3gp video) using a secure HTTP URL (one that begins with "https"), then you must install the security certificate from the server that holds the media file on your Voice Portal system. To install the security certificate on Voice Portal, go to the Certificates page of the VPMS and select the Trusted Certificates tab.
As a matter of security policy, Avaya recommends that the Linux service named portmap be disabled. By default, the portmap service is disabled on Avaya Linux. However, on Red Hat Enterprise Linux 5.x the portmap service is enabled by default.
To disable the portmap service:
chkconfig portmap off |
/sbin/service portmap stop |
Note: The Voice Portal backup procedure requires you to have the portmap service enabled. Similarly, you will need to enable the portmap service if you wish to use NFS.
In Voice Portal 5.1 if you have multiple SIP proxies configured and an outcall request fails because the chosen SIP proxy is out of service, Voice Portal will automatically retry the outcall using a different SIP proxy. However, it may take Voice Portal up to thirty seconds to retry the outcall using a different SIP proxy. Writers of applications that request outcalls using either the LaunchVXML method of the Application Interface web service or the CCXML tag <createcall> should consider this failover time when choosing a timeout to specify on the outcall request. If the timeout specified is too short, the outcall request will time out before Voice Portal has found a working SIP proxy.
Note: If a LaunchVXML request times out, it is still possible that Voice Portal will eventually complete the outcall and launch the requested application.
If your application makes an outcall with call classification enabled in either a SIP environment or an H.323 environment that has Special Application 8874 (also known as "the green feature") enabled, then the call classification timer does not begin until Voice Portal receives an indication from the telephony switch that the far end has answered the call. Therefore, if a call goes unanswered, the application will not receive a call classification timeout. It will receive a connection timeout instead.
In an H.323 environment that does not have Special Application 8874 enabled, on the other hand, Voice Portal does not receive any indication from the telephony switch when the far end has answered the call. In this case, Voice Portal considers the call connected as soon as the call is initiated and Voice Portal immediately starts the call classification timer. If a call goes unanswered, your application will not receive a connection timeout. It will receive a call classification timeout.
If the host name or IP address of any Voice Portal server changes, then you must run the utility do_UpdateHost.
If the host name or IP address of a Primary VPMS or Auxiliary VPMS server changes, you must additionally perform the following steps on that server after running doUpdateHost:
/sbin/service vpms stop |
cp -p /tmp/BootstrapDBResource.dat /opt/coreservices/scc/runtime/ |
/sbin/service vpms start |
To replace a Primary VPMS server you must perform several tasks that include: install Linux on the new server, install Voice Portal on the new server, and restore on the new server a Voice Portal database backup taken from the old server. Below are some additional procedures that need to be followed during this process. The first procedure should be performed before you restore the Voice Portal database backup on the new newer. The second procedure should be performed after you restore the database.
Important: These additional procedures are needed any time a database restore is performed after a fresh install of Voice Portal; even if the underlying server hardware has not changed.
Before following the procedure documented in Restoring data backed up from System Backup follow the procedure below.
For each Auxiliary VPMS server:
cd $AVAYA_HOME/Support/VP-Tools |
bash setup_vpms.php <PrimaryVPMS> |
Note: The <PrimaryVPMS> above is the name or IP address of your Primary VPMS server.
After following the procedure documented in Restoring data backed up from System Backup follow the procedure below.
/sbin/service vpms restart |
Under rare circumstances, a Voice Portal MPP will generate a core dump after you make a change to the network configuration on one of your speech servers. If you make a network configuration change on a speech server, such as changing the server's IP address, reboot the speech server. This should prevent the situation that causes the MPP to core dump.
Voice Portal communicates with speech servers using either Media Resource Control Protocol version one (MRCPv1) or Media Resource Control Protocol version two (MRCPv2). A variety of issues, including failure to acquire speech resources and long latencies, have occasionally been observed when Voice Portal uses MRCPv2 to communicate with speech servers. If you experience problems using MRCPv2, configure Voice Portal to use MRCPv1 instead.
Note: Problems with MRCPv2 are particularly common when using Loquendo Speech Servers. You should always configure Voice Portal to use MRCPv1 when communicating with Loquendo Speech Servers.
If you are running a Dialog Designer application that does application logging, and your Voice Portal system is configured to handle more than 75 simultaneous calls, you must make the following configuration change on your Primary VPMS:
/sbin/service vpms restart |
When an application runs on the MPP, the VoiceXML Interpreter defaults to the language configured in the Language field on the AVB Settings page of the VPMS. The default value for this field is en-US. If your application uses a language other than the one configured for the system, the application must specify the correct language using the xml:lang attribute. Applications can specify the language either globally on each page, or as part of each individual <grammar> tag and <prompt> tag.
When the VoiceXML Interpreter or CCXML Interpreter generates an event to a page that does not have a handler defined for that event, the interpreter invokes the appropriate default event handler configured on the Event Handlers page of the VPMS. The default event handlers shipped with Voice Portal play an error prompt spoken in English. If your application uses a language other than English, you must configure a default event handler appropriate for that application.
Note: If you are writing code that will be used as a default event handler and the code refers to prerecorded prompt files, refer to those prompt files by file name as opposed to file path. Voice Portal will automatically search for prompt files in the appropriate directory.
If you are running a Dialog Designer application that does application logging, and your VPMS goes out of service, Dialog Designer temporarily stores the application log data on your application server while waiting for the VPMS to come back in service. If your application server is Tomcat running on Linux and the VPMS stays out of service for a long time, Tomcat might crash because the system runs out of file handles.
If you encounter this problem, do one or both of the following on your application server.
Increase the Linux file handle limit on your application server:
ulimit -n 8192 |
/sbin/service tomcat restart |
Change Tomcat startup options to improve garbage collection on your application server:
export JAVA_OPTS="-server -Xmx1024m -XX:MaxNewSize=30m -X:+UseParNewGC -XX:+UseConcMarkSweepGC -XX:CMSInitiatingOccupancyFraction=60-XX:ThreadStackSize=512" |
/sbin/service tomcat restart |
In order for Voice Portal to establish a TLS connection to a Session Manager server, you must install the security certificate from Session Manager on Voice Portal. The correct URL to obtain the security certificate from Session Manager is:
https://<Session_Manager_SM-100>:5061
Note: The <Session_Manager_SM-100> above is the name or IP address of the SM-100 asset within your Session Manager server.
In order for Voice Portal to establish a TLS connection to a Session Manager server, you must install the security certificate from Voice Portal on Session Manager. Please contact Avaya technical support for help with installing the Voice Portal security certificate on Session Manager.
In order for Voice Portal to establish a TLS connection to a SIP Enablement Services (SES) server, you must install the security certificate from SES on Voice Portal. The correct URL to obtain the security certificate from SES is:
https://<SES_Server>:5061
Note: The <SES_Server> above is the name or IP address of your SES server.
When all the MPPs in your system are in service, the SES properly balances the call load among the MPPs. If one MPP is taken out of service, though, all calls originally intended for the out-of-service MPP are rolled over to one of the remaining MPPs rather than being spread evenly among all of the remaining MPPs.
Note: Generally, this load imbalance will not cause any problems with the remaining MPPs taking calls. However, if your application makes any outcalls, such as during a bridged transfer, then it may encounter no-resource errors. These errors will occur if the MPP taking additional load from the out of service MPP has all ports busy.
Under some conditions, an MPP server may consume ASR/TTS resources on a Loquendo speech server even when there are no calls active on that MPP. This can prevent other MPP servers that share the same Loquendo speech server from being able to obtain speech resources when they are processing a call.
To work around this issue:
You might see the error "A security problem occurred" from the Windows Media Player while playing an utterance from a Session Detail report. To listen to the utterance, right-click the utterance link and select Save Target As. For more information, see http://support.microsoft.com/default.aspx?scid=kb;en-us;885136.
When you click the Export button on any VPMS page that includes this option, a File Download dialog box containing Open, Save, and Cancel buttons is displayed. If the Internet Explorer security option Do not save encrypted pages to disk is enabled, the Open button will not function properly.
To work around this issue, either:
Or
While viewing an e-mail that was sent by Voice Portal as the result of a scheduled report running, if you click the link to view the report output you may see either a blank page or the following message:
The requested file is no longer available.
This problem has only been observed when the scheduled report name contains multi-byte characters and the browser used is Internet Explorer 6.0.
To work around this issue, either:
Or
If you are using Nuance speech servers and you have any applications with the Advanced Parameter Support Remote DTMF Processing set to Yes, then your Nuance speech servers must all be running NSS 5.0.4 or later. This is to prevent <noinput> VoiceXML exceptions from occurring for remote DTMF detection.
If your application plays a long prompt with barge-in enabled, the Nuance speech server may return a recognition timeout event before the prompt has completed playing. To remedy this problem, increase the session timeout parameter on the Nuance speech server to be longer than your longest application prompt.
For a Nuance Speech Server running on Linux:
server.mrcp2.sip.sessionTimeout VXIInteger XXXXXX |
Note: The XXXXXX above is the session timeout in milliseconds. For example, for a two minute timeout XXXXXX would be 120000.
server.mrcp1.rtsp.sessionTimeout VXIInteger XXXXXX |
Note: The XXXXXX above is the session timeout in milliseconds. For example, for a two minute timeout XXXXXX would be 120000.
For a Nuance Speech Server running on Windows:
server.mrcp2.sip.sessionTimeout VXIInteger XXXXXX |
Note: The XXXXXX above is the session timeout in milliseconds. For example, for a two minute timeout XXXXXX would be 120000.
server.mrcp1.rtsp.sessionTimeout VXIInteger XXXXXX |
Note: The XXXXXX above is the session timeout in milliseconds. For example, for a two minute timeout XXXXXX would be 120000.
If you are using Nuance speech servers in their default configuration, when your application receives a <nomatch> event in response to a recognition request the application variable application.lastresult$ will not be updated with any recognition results. If you want your application to receive recognition results when a <nomatch> event is generated, then your Nuance speech servers must all be running NSS 5.0.7 or later and they must all be running NRec 9.0.11 or later. Additionally, you must perform the procedure below on each of your Nuance speech servers.
For a Nuance Speech Server running on Linux:
server.mrcp2.osrspeechrecog.mrcpdefaults.VSP.server.osrspeechrecog.result.sendnomatch VXIString true |
Note: If present, be sure to remove the # character at the beginning of the above line to uncomment it.
server.mrcp1.osrspeechrecog.result.sendnomatch VXIString true |
Note: If present, be sure to remove the # character at the beginning of the above line to uncomment it.
<param name="swisr_result_enable_speech_mode"> |
<param name="swisr_result_enable_speech_mode"> |
<value>1</value> |
</param> |
<param name="swirec_result_enable_speech_mode"> |
<param name="swirec_result_enable_speech_mode"> |
<value>1</value> |
</param> |
For a Nuance Speech Server running on Windows:
server.mrcp2.osrspeechrecog.mrcpdefaults.VSP.server.osrspeechrecog.result.sendnomatch VXIString true |
Note: If present, be sure to remove the # character at the beginning of the above line to uncomment it.
server.mrcp1.osrspeechrecog.result.sendnomatch VXIString true |
Note: If present, be sure to remove the # character at the beginning of the above line to uncomment it.
<param name="swisr_result_enable_speech_mode"> |
<param name="swisr_result_enable_speech_mode"> |
<value>1</value> |
</param> |
<param name="swirec_result_enable_speech_mode"> |
<param name="swirec_result_enable_speech_mode"> |
<value>1</value> |
</param> |
In order to generate either Traditional Chinese or Simplified Chinese TTS using Nuance RealSpeak 4.0.10, you must configure RealSpeak correctly and ensure that the application behaves properly.
To generate Traditional Chinese:
server.realspeak4.language.XX.ShortName VXIString zh-guoyu |
Note: The XX above is a number.
server.realspeak4.language.XX.ShortName VXIString zh-CN |
Note: The XX above is a number.
server.realspeak4.language.XX.ShortName VXIString zh-guoyu |
Note: The XX above is a number.
server.realspeak4.language.XX.ShortName VXIString zh-CN |
server.realspeak4.language.XX.FullName VXIString Mandarin Chinese Big5 |
server.realspeak4.language.XX.Voice VXIString Mei-Ling |
server.realspeak4.language.XX.Gender VXIString female |
Note: The XX above is a number.
To generate Simplified Chinese:
server.realspeak4.language.XX.ShortName VXIString zh-guoyu |
Note: The XX above is a number.
server.realspeak4.language.XX.ShortName VXIString zh-CN |
Note: The XX above is a number.
server.realspeak4.language.XX.ShortName VXIString zh-guoyu |
Note: The XX above is a number.
server.realspeak4.language.XX.ShortName VXIString zh-CN |
server.realspeak4.language.XX.FullName VXIString Mandarin Chinese GB |
server.realspeak4.language.XX.Voice VXIString Mei-Ling |
server.realspeak4.language.XX.Gender VXIString female |
Note: The XX above is a number.
In order to generate either Traditional Chinese or Simplified Chinese TTS using Nuance RealSpeak voice Mei-Ling 4.0.6 with Nuance Speech Server (NSS) 5.0.x, you must configure NSS correctly.
For a Nuance Speech Server running on Linux:
server.rsspeechsynth.languageid.zh VXIString Mandarin Chinese |
server.rsspeechsynth.languageid.zh-CN VXIString Mandarin Chinese |
server.rsspeechsynth.languageid.zh-guoyu VXIString Mandarin Chinese |
server.rsspeechsynth.languageid.zh VXIString Mandarin Chinese |
server.rsspeechsynth.languageid.zh-CN VXIString Mandarin Chinese GB |
server.rsspeechsynth.languageid.zh-guoyu VXIString Mandarin Chinese B5 |
For a Nuance Speech Server running on Windows:
[HKEY_LOCAL_MACHINE\SOFTWARE\ScanSoft\RealSpeak 4.0\Language Resources\Mandarin Chinese B5 (Mei-Ling)] |
"Gender"="female" |
"LanguageName"="Mandarin Chinese B5" |
"LanguageTag"="zh-guoyu-b5" |
"VoiceName"="Mei-Ling" |
[HKEY_LOCAL_MACHINE\SOFTWARE\ScanSoft\RealSpeak 4.0\Language Resources\Mandarin Chinese GB (Mei-Ling)] |
"Gender"="female" |
"LanguageName"="Mandarin Chinese GB" |
"LanguageTag"="zh-guoyu-gb" |
"VoiceName"="Mei-Ling" |
server.rsspeechsynth.languageid.zh VXIString Mandarin Chinese |
server.rsspeechsynth.languageid.zh-CN VXIString Mandarin Chinese |
server.rsspeechsynth.languageid.zh-guoyu VXIString Mandarin Chinese |
server.rsspeechsynth.languageid.zh VXIString Mandarin Chinese |
server.rsspeechsynth.languageid.zh-CN VXIString Mandarin Chinese GB |
server.rsspeechsynth.languageid.zh-guoyu VXIString Mandarin Chinese B5 |
More Information
Online: