> For the complete documentation index, see [llms.txt](https://releases.holoplot.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://releases.holoplot.com/october-2026/holoplot-os-2.6.3.md). # HOLOPLOT OS 2.6.3 We're excited to announce **HOLOPLOT OS 2.6.3**. This release focuses on quality, stability, monitoring and traceability. It addresses issues and residual risks of the previous OS version, especially those related to single-Controller dependencies, which caused transient Processor interruptions and HOLOPLOT Control downtime. Automatic Processor failover is now bidirectional and health-aware. {% hint style="info" %} The Public System API has been updated to **v1.7.0**. It adds an endpoint to enable or disable Show Mode from a venue control system and extends the Show Mode status endpoint. All previously existing endpoints are unchanged. {% endhint %} *** ### What's New #### Health-aware, bidirectional Processor failover [Automatic Processor failover](https://docs.holoplot.com/holoplot-control/maintaining-the-system/processor-failover) now works from Main to Backup and from Backup to Main, in each case only when the target Processor is healthy: * If the Main Processor fails and the Backup Processor is healthy, the Backup Processor takes over automatically. * If the Backup Processor is offline or in an error state, the automatic failover is blocked. HOLOPLOT Control shows an alert and the incident is logged in the [Event Log](https://docs.holoplot.com/holoplot-control/maintaining-the-system/device-events-list). A manual failover remains available. * If the Backup Processor fails while it is active and the Main Processor is healthy, the system switches back to the Main Processor automatically. #### Remote Processor reboot Restart a Processor, including its Audio Bridge, from the device list or the Processor side panel in HOLOPLOT Control instead of power-cycling it on site. Available for online Processors, whether paired or not. Disabled while [Show Mode](https://docs.holoplot.com/holoplot-control/operating-the-system/show-mode) is active. #### Clearer device status * Unified [Online, Offline and Sleeping](https://docs.holoplot.com/holoplot-control/maintaining-the-system/monitoring-device-health#device-status) badges, counters and filters across the device list, network settings, outputs page, footer and side panels. * Arrays show how many of their expected modules are paired. * Modules that are paired in the project but have never been seen online are shown in the device list and the pairing view as "Never seen online". * The Processor side panel shows the Audio Bridge ID. * Modules with a PTP clock fault are flagged in the health view of the side panel. * Network groups show their device count. * Bulk actions keep the current selection and show how many modules they will affect. #### Redesigned metering [Output meters](https://docs.holoplot.com/holoplot-control/operating-the-system/monitoring-level-meters#system-array-and-module-output-metering) use a continuous green-to-red gradient anchored to real dB levels, so brief peaks near 0 dB no longer turn the whole bar red. Meters stay visible but dimmed while the system is muted or asleep. #### Show Mode in the Public System API * `PUT /show-mode` enables or disables [Show Mode](https://docs.holoplot.com/holoplot-control/operating-the-system/show-mode) from the venue control system. * `GET /status/show-mode` also reports whether Audio Macros and Macro snapshots are allowed in Show Mode. *** ### Improvements and bug fixes #### HOLOPLOT Control * Faster 3D viewport rendering in projects with many modules. * Fixed the Audio Bridge PTP clock domain being displayed as "--". * Fixed restoring a Settings snapshot failing with a validation error. * Fixed the Settings snapshot dialog offering Gain and Delay for function groups, which do not support them. * Fixed meters blinking or showing incorrect values in some Chrome versions. * Fixed Processor health status occasionally failing to load. * Fixed the Devices and Module Pairing pages in rare cases showing modules as unpaired and sleeping while they were actually offline or removed. * Fixed the Device Summary sometimes not updating when a Controller or Processor went offline or came back online. Redundant Processor pairs are now counted as two devices, matching the device list. * A missing HOLOPLOT VPN connection no longer counts against [system health](https://docs.holoplot.com/holoplot-control/maintaining-the-system/monitoring-device-health). The VPN is consistently named "HOLOPLOT VPN" in the Event Log, the issue list and the side panels. * The read-only Project Information view stays reachable while Show Mode is on. * Disabled Macro knobs are hidden from the Sections properties panel. * Removed the connection status indicator from the [footer](https://docs.holoplot.com/holoplot-control/getting-started/navigating-the-interface#footer). * Redesigned the Unmute and Unsolo Beams buttons in the sidebar for better visibility. * The message for an unhealthy Controller or Processor power supply now reads "Not healthy" instead of "Above expected voltage", matching the new iDRAC-based power supply health reporting. * Password-protected settings and tuning pages no longer reload on every navigation, which caused frequent "connecting to controller" screens. * Fixed various smaller UI issues: misaligned meters in the array header, long module IDs not fitting the metering card, the stream settings dialog being too small for the stream list, and the text color of disabled buttons. #### System Backend * Improved Controller cluster resilience when a single Controller is lost or rebooted, for better HOLOPLOT Control uptime and faster reconnection. * Every Controller keeps a local name-resolution (DNS) cache, removing the name-resolution failures seen on the remaining Controllers during a Controller reboot. * A Controller that cannot be reached is now reported offline in HOLOPLOT Control instead of staying online, except in the case described in "Monitoring temporarily unreliable after a Controller goes offline" under [Known issues](https://docs.holoplot.com/user-guides/troubleshooting#known-issues)). * Fixed Controllers occasionally renaming themselves (for example from RPF-9654 to RPF-9655) after a network link flap, which made them unreachable by name. * Fixed spurious "Controller offline" events in the Event Log. * Fixed spurious fan health events in the Event Log. * Controller and Processor power supply health is now based on the iDRAC's power supply health data point instead of a single input-voltage threshold. This addresses spurious power supply failure events in the Event Log. A disconnected power supply no longer shows an impossible negative voltage. * Network configuration changes and snapshot restores now affect only modules in the active project. * Fixed importing a project resetting the network configuration of every module in the venue, including modules of other projects and modules with manually assigned networks. * Fixed restoring a system settings snapshot failing because of conflicting network settings between the snapshot and the live system. * Fixed historical events from before the PTP terminology change failing to load in the Event Log. * Improved cloud telemetry: system events, such as failovers and configuration changes, are reported in the same structured format as device events, and device logs are no longer rate-limited, so nothing is dropped during an incident. #### HOLOPLOT Processor * Improved Processor resilience when a Controller in the cluster is lost. This prevents the erroneous Processor failovers and Processor audio interruptions seen in the previous version, and fixes an issue where the Processor did not notice a lost connection to a Controller after a network interruption, which could delay a needed failover or leave streams running. * Fixed several race conditions in the Processor software that could cause a crash during reconnects. * Fixed a Processor not restoring its audio configuration after it reconnected to the Controllers, which could leave its streams unconfigured after a network interruption. * Tuned Controller disconnect detection and failover delays for the Processors and the failover service to minimize audio interruption during a failover. * Fixed metering messages piling up while a Processor was disconnected. * Fixed Processor network interface names not always being applied after a cold boot. * Audio streams that the Processor sets up on the Audio Bridge are removed automatically when the Processor software disconnects. Audio streams created manually in the Audio Bridge web UI are kept. Processor transmit streams are no longer listed among the discovered third-party streams, and receive streams are activated once the Audio Bridge's PTP is locked. * Mitigated silent Audio Bridge streams after an Audio Bridge software restart: the Processor software now detects this condition and restarts itself so that the streams are rebuilt. * Processors report the full iDRAC system health to cloud telemetry, like Controllers, and the iDRAC is polled less often to avoid sensor-related issues. #### Audio Modules * Improved Audio Module resilience to short interruptions of the connection to the Controller cluster. * Fixed Audio Modules not reacting to network link changes in rare cases. * A missing IP address on the secondary audio network interface of an X1 module in a redundant network setup is now reported as a [health issue](https://docs.holoplot.com/holoplot-control/maintaining-the-system/device-issues-list#network-issues). *** ### Known Issues Known issues in HOLOPLOT OS 2.6.3, their workarounds and their status are listed in [Known issues](https://docs.holoplot.com/user-guides/troubleshooting#known-issues) on the Troubleshooting page.