Troubleshooting the Pexip Infinity platform

This topic contains a list of symptoms, possible causes and suggested resolutions for some issues you may experience when using the Pexip Infinity platform.

It includes the following sections:

Additional, separate, help pages are available for other integration scenarios:

For an up-to-date list of devices that are supported by Pexip Infinity, including any known issues, see Interoperability.

Contacting support

If you cannot find the information you require, contact your Pexip authorized support representative. Technical support for software issues is available while under a valid support contract and running a Pexip Infinity version no more than 2 major software releases behind the current release. Software bug fixes will only be provided in either the current or the next major release of software.

Pexip Infinity deployment and upgrading

Symptom Possible cause Resolution
During upgrade, one or more Conferencing Nodes are stuck with a status of "waiting for calls to clear", but there are no active calls reported on the Management Node. A completed call has not cleared properly from the Conferencing Node. Reboot the Conferencing Node.
A Conferencing Node does not accept calls even though it is powered on and is contactable on the network.

If time is not properly synchronized between the Management Node and the host server, certificates issued by the Management Node may be invalidated by Conferencing Nodes within the same Pexip Infinity deployment. As a result, the Conferencing Nodes will not communicate properly with the Management Node, causing calls to fail.

Ensure all virtual machines (i.e. the Management Node and all Conferencing Nodes) within the Pexip Infinity platform, and the host servers on which they are running, are using accurate times according to the public or private standard NTP clock.

We strongly recommend that you configure at least 3 distinct NTP servers or NTP server pools in each instance to ensure proper synchronization.

To synchronize time on Pexip Infinity:

  1. Synchronize time on the host servers (for instructions, see the relevant hypervisor installation guide).
  2. Enable NTP on Management Node.
  3. Reboot all VMs.

A newly deployed Conferencing Node does not accept calls and its last contacted status on the Management Node shows "Never", even though it is powered on and is contactable on the network.

After deploying a new Conferencing Node, it takes approximately 5 minutes before the node is available for conference hosting and for its status to be updated on the Management Node. Until it becomes available, the Management Node reports the status of the Conferencing Node as having a last contacted and last updated date of "Never". "Connectivity lost between nodes" alarms relating to that node may also appear temporarily.

Wait for the Conferencing Node to finish initializing.
The wrong information was entered while running the installation wizard.  

On server-based deployments you can re-run the installation wizard by following the instructions in Re-running the installation wizard.

Do not re-run the installation wizard on cloud-based deployments (Azure, AWS, GCP or Oracle) in order to change Management Node configuration data such as its IP address or hostname. To change such data you must terminate the existing instance and deploy a new Management Node instance. You should only re-run the installation wizard on cloud-based deployments if you need to reset the web administration password (and then you should not change any of the other configuration data).

A new Management Node or Conferencing Node does not work. It was created by cloning it through VMware. You cannot use cloning to create Management Nodes or Conferencing Nodes.

Create the Management Node according to our instructions (see Pexip Infinity installation overview ).

Create all Conferencing Nodes by following the instructions in Deploying new Conferencing Nodes.

A newly-deployed Conferencing Node has gone into maintenance mode with the message "CPU instruction set is not supported; system will be placed in maintenance mode". The Conferencing Node has been installed on a system that does not meet the CPU instruction set requirements.

Do one of the following:

Dynamic bursting to a cloud service

Symptom Possible cause Resolution
No bursting nodes appear in the Cloud overflow nodes area of the Status > Cloud bursting page The cloud node instances are not tagged correctly. Check that the node instances running in your service provider have been assigned the pexip-cloud tag and that the tag value is set to the Management Node hostname.
You see a status issue "Instance <name> (with IP <address>) was found, but no corresponding Conferencing Node has been configured".

This occurs when Pexip Infinity detects a bursting instance with a tag matching your system's hostname but there is no corresponding Conferencing Node configured within Pexip Infinity.

This message can occur temporarily in a normal scenario when deploying a new Conferencing Node and you have set up the VM instance in your cloud provider but you have not yet deployed the Conferencing Node in Pexip Infinity. In this case, the issue will disappear as soon as the Conferencing Node is deployed.

You see connectivity errors in the administrator log while overflow nodes are being started/stopped. This is normal behavior. No action required.
"Not authorized to perform an operation on <instance ID or region name>. Check the policy created for the AWS user." error. This means that there is a problem with the AWS policy document, or the AWS user is not attached to the policy. See Configuring an AWS user and policy for controlling overflow nodes for more information.
You see a status issue "Cloud bursting process encountered the following error: unsupported operand type(s) for -: 'datetime.datetime' and 'NoneType'" A system location's configured Transcoding location only contains bursting nodes. Ensure that the system location's configured Transcoding location contains "always-on" nodes.

Joining a conference and viewing content

Symptom Possible cause Resolution

Participants cannot join a conference due to insufficient capacity.

  • When users attempt to join a conference they get a message saying "Participants cannot join a conference due to insufficient capacity."
  • There is an alarm "Call capacity limit reached".
  • The administrator log is reporting "Participant failed to join conference" and "out of proxying resource" or "out of transcoding resource".

A call has not been accepted because all Conferencing Nodes that are able to take the media for this call are at capacity. It could be either Proxying Edge Nodes or Transcoding Conferencing Nodes that are out of capacity.

  • Deploy more Conferencing Nodes in either the proxying or transcoding location as appropriate.
  • Move existing Conferencing Nodes onto more powerful servers.
  • Allocate more virtual CPUs for Conferencing Nodes on existing servers (if there are sufficient CPU cores). Note that the Conferencing Node will have to be rebooted for this to take effect.
  • Configure each location with a primary and secondary overflow location.
  • If a call is received in a location that contains Proxying Edge Nodes, that location must be configured with a Transcoding location that contains your Transcoding Conferencing Nodes.

Note that some types of call consume more resources than other calls. Thus, for example, if you are at full capacity and an audio-only call disconnects, there may still not be sufficient free resource to connect a new HD video call. For further information on capacity and how calls consume resources, see Hardware resource allocation rules.

The participant has dialed in to the Conferencing Node while it is still starting up and an internal capacity-checking tool is running. Wait for one minute and then attempt to join the conference.
The Virtual Meeting Room or Virtual Auditorium has a participant limit applied, and this limit has been reached. Increase the participant limit, if appropriate.

Participants cannot join a conference due to an invalid license.

  • When users attempt to join a conference they get a message saying "Participants cannot join a conference due to an invalid license."
  • The admin log is reporting "Participant failed to join conference" and "no valid license available".

If your Pexip Infinity reports an invalid license, this could mean that:

  • the license has not been activated
  • the existing license has expired
  • the existing license has become corrupt (this could occur, for example, if the Management Node reboots after an upgrade and comes back up on a different physical blade with a new MAC address).
  • Check the status of your licenses from the Licensing page (Platform > Licenses).
  • Contact your Pexip authorized support representative for assistance.

For more information, see Pexip Infinity license installation and usage.

Participants cannot join a conference due to insufficient licenses.

  • When users attempt to join a conference they get a message saying "Participants cannot join a conference due to insufficient licenses."
  • There is an alarm saying "License limit reached".
  • The admin log is reporting "Participant failed to join conference" and "license limit reached".
There are not enough call licenses available on the system at this time. For more information, see Pexip Infinity license installation and usage.
  • Wait until one or more of the existing conferences have finished and the call licenses have been returned to the pool.
  • Contact your Pexip authorized support representative to purchase more call licenses.
An H.323 endpoint has its bandwidth restricted when joining a conference via a Virtual Reception, or when placing a call using the Distributed Gateway after first connecting to a Virtual Reception. If there has been a bandwidth restriction placed on the Virtual Reception, any H.323 endpoints using that service will not be able to subsequently increase their bandwidth, even after being transferred to a Virtual Meeting Room or using a Call Routing Rule that has a higher (or no) limit.
  • Make the call using a SIP endpoint or Connect app.
  • Do not place a bandwidth restriction on the Virtual Reception.
Presentations do not display full screen.

If the presentation being shared is either:

  • an application that is not in full-screen mode
  • a full screen image that is being sent from a non-standard aspect ratio screen

then the image being sent may have a non-standard aspect ratio. To send the image inside a standard resolution window (for example 640x480 [4:3]) or 1280x720 [16:9]), the endpoint may add horizontal or vertical mattes (known as letterboxing or pillarboxing respectively).

Ensure that presenters always either:

  • share their entire screen, or
  • share individual applications when they are in full-screen mode only.

Images are not displaying as expected:

  • they are being cropped
  • they have black bars at the top or sides

Endpoints send and display video images and presentations in various aspect ratios, most commonly 16:9 and 4:3. If there is a difference between the aspect ratios of the sending and receiving endpoints, then the endpoint and/or Pexip Infinity may crop the image or add vertical or horizontal mattes.

For more information, see Changing aspect ratios.

Occasional video freezes are seen in media received from SIP or H323 endpoints, where Conferencing Nodes in GCP are in use. System locations that include Conferencing Nodes running in Google Cloud Platform must be configured with a maximum MTU of 1460 bytes. Some older endpoints default to a MTU higher than this. Reduce the MTU on the endpoint to 1350.

Main video on a Cisco E20 freezes and no presentation is shown when a VMR participant starts presenting.

The E20 still sends video as normal, and audio flows in both directions.

This may occur when additional video codecs (e.g. H.264 High Profile) are enabled on the Pexip Infinity deployment, and if the E20 is called from a VMR and then a participant in the VMR presents content.

This occurs because more codecs are offered to the E20 than it can cope with.

Disable all video codecs (i.e. H.264 High Profile) not required in the deployment (Platform > Global settings > Codecs).

Cisco endpoints running TC version 5.x or earlier software do not receive video when dialed out to from Pexip Infinity. Video is still sent as normal.

This may occur when additional video codecs (e.g. H.264 High Profile) are enabled on the Pexip Infinity deployment, and the TC5.x or older is called from a VMR.

This occurs because more codecs are offered to the TC endpoint than it can cope with.

  • Upgrade the TC5.x endpoint to a more recent software version.
  • Disable all video codecs (i.e. H.264 High Profile) not required in the deployment (Platform > Global settings > Codecs).
A Cisco endpoint running TC7.1.4 might not receive video from Pexip in H.323 calls. This is due to a bug in TC7.1.4 software.
  • Disconnect and reconnect the call.
  • Use SIP instead of H.323.
  • Upgrade the TC software.
Cisco endpoints running TC7.0.x or earlier may crash when connecting to or from Pexip Infinity v24 or later over SIP. This happens when AES-256 cipher is included in the SDP. Upgrade the TC software to the latest version.
Poly endpoint is displaying its SIP address instead of its System Name in calls. The endpoint is configured to use its SIP address.

To configure your Polycom device to display its system name in calls instead of its SIP address:

  1. Go to General Settings > System Settings and set a Device name (remember to press Save at the bottom).
  2. Go to Call Configuration > Call Settings and select the Display System Name Instead of SIP Address option.

Note that some older software versions do not support these options — the minimum requirements are:

  • RealPresence Group requires 6.2.2 or later.
  • Studio X Series requires 3.13.0 or later.
Participants keep hearing themselves repeated back after a short delay. This happens in a conference with one or more other participants connected using the Connect web app via Edge, Safari or Firefox, and who are using their computer's microphone and speakers. Edge, Safari and Firefox do not have adequate echo cancellation, and in certain circumstances may experience a delay in playing audio. When this happens, sounds played through the computer's speakers are picked up by the computer's microphone and replayed back to other participants.

Participants using Edge, Safari or Firefox should:

  • use a headset
  • mute themselves when not speaking
  • consider using Chrome or the Connect desktop app instead.
In-band DTMF tones may not be detected if they are input too quickly. False detections can be caused, for example, by poor line quality, line noise and echo. This is best resolved through using out-of-band DTMF tones.
Participants are disconnected from conferences and "Backplane disconnected" messages are recorded in the administrator log.

VMware snapshots were being taken or deleted while conferences were in progress. Taking or removing snapshots can cause virtual machines to become unresponsive.

For more information, see this VMware knowledge base article.

Only create and delete VMware snapshots at a time of minimal usage.

Conference connectivity and TLS issues

Symptom Possible cause Resolution
Calls fail when dialing out to, or receiving calls from, video network infrastructure devices, such as a Cisco VCS or CUCM. Calls may fail immediately, or after a period of time, and SSL alerts are raised in the support log with an "unsupported certificate" description. Either the external system or the Conferencing Node is verifying the other party's certificate and it is rejecting the connection because the certificate does not have client authentication properties. Ensure that the certificates on your external systems and on your Conferencing Nodes contain "TLS Web Client Authentication" Enhanced Key Usage properties (see Mutual TLS authentication and client/server certificates).
A Cisco MXP intermittently puts a call on hold immediately after resuming it. This is due to a bug in the MXP where a race condition exists between the resume message and the session refresh re-INVITE message. Set the session refresh configuration on your call control system to a value that avoids this race occurring.

TLS certificate administration

Error message Possible cause Resolution
Certificate and private key do not appear to be part of the same key pair This most likely means that you have tried to upload the certificate against the wrong CSR. Select the correct CSR and try again.

Pexip Infinity administration

Symptom Possible cause Resolution
A policy profile is configured but it is being ignored. The policy profile has not been assigned to a system location. Assign the policy profile to your locations (Platform > Locations).

Log timestamps appear to be inaccurate or log entries appear to be out of sequence.

Time is not properly synchronized between the Management Node, Conferencing Nodes and their host servers, causing different systems to use different timestamps. This could be because:

  • Insufficient NTP servers or NTP pools have been configured on a host server or the Management Node (we recommend a minimum of 3).
  • One or more NTP servers are unreachable or have inaccurate time themselves.
  • NTP is otherwise not configured according to our recommendations.

Ensure all virtual machines (i.e. the Management Node and all Conferencing Nodes) within the Pexip Infinity platform, and the host servers on which they are running, are using accurate times according to the public or private standard NTP clock.

We strongly recommend that you configure at least 3 distinct NTP servers or NTP server pools in each instance to ensure proper synchronization.

To synchronize time on Pexip Infinity:

  1. Synchronize time on the host servers (for instructions, see the relevant hypervisor installation guide).
  2. Enable NTP on Management Node.
  3. Reboot all VMs.

How do I register Pexip Infinity to a gatekeeper?   You don't register Pexip Infinity to a gatekeeper. Instead, configure your call control system to route calls to Pexip Infinity. See Call control.
Oracle Acme Packet SBC has the error "Message Too Large".  

Increase the following configuration parameters on the Acme Packet SBC:

  • sip-message-len 16000
  • option +max-udp-length=0
VMware datastore is showing disk I/O alarms.  

Enabling SIOC on your datastores might help. For more information, see this VMware knowledge base article.

A participant's Display name has some characters replaced with asterisks. For H.323 endpoints, the display name should use IA5 characters, but some endpoints use UTF-8 or other encodings. If Pexip Infinity receives any non-IA5 characters, it replaces them with an asterisk. Ensure the display name used by such endpoints uses a string that contains valid IA5 characters only.
Unable to take snapshots on deployments hosted in Microsoft Azure. The Azure Load Balancer is timing out. Increase the TCP idle timeout setting in Azure for the Azure Load Balancer. See https://docs.microsoft.com/en-us/azure/load-balancer/load-balancer-tcp-idle-timeout for configuration instructions.

Connect apps

The table below describes general issues that may occur when using the Connect apps within your deployment. For a list of specific error messages that may be presented to Connect app users, along with their meaning and suggested resolution (where appropriate), see Troubleshooting Connect app error messages.

Symptom Possible cause Resolution
Participants cannot use Connect apps. Connect web app and Connect desktop app users are presented with a message "This feature has been disabled." Support for these Connect apps has been disabled.

Enable support as follows:

  1. Go to Platform > Global settings > Connectivity.
  2. Select the Enable support for Pexip Infinity Connect clients and Client API checkbox.
Participants attempting to connect to a conference using the Connect mobile app for iOS get a message "Certificate Error. Please contact your system administrator." Your deployment does not use valid, trusted certificates. Connect mobile apps require deployments with HTTPS and valid certificates. Install valid certificates. For more information, see Managing TLS and trusted CA certificates.
Participants using a Connect app cannot write or see chat messages. Chat has been globally disabled. Enable chat.
Guest participants using a Connect app cannot see the chat window, participant list, or presentation. Guest participants are not allowed in to the conference if the conference is locked, or until the first Host participant has joined.
Users of the Connect web app see "An error occurred. The page you are looking for is currently unavailable." error message when connecting via a reverse proxy. You are running version 3 or earlier of the Pexip Reverse Proxy and TURN Server and the web browser cannot make a TLS connection. Upgrade to latest version of the Pexip Reverse Proxy and TURN Server.

Users see a Connection Alert message at the top of their window

There is packet loss of 3% or more between the Connect app and Pexip Infinity. Users who experience this issue regularly should reduce their bandwith by selecting a lower Connection Quality.