AVEVA Releases System Platform 2023 R2 SP1 Patch 02

AVEVA released System Platform 2023 R2 SP1 Patch 02 (P02) in July 2025. I’ve gone through every ReadMe file in the package so you don’t have to. Here’s my summary of what’s in this patch, what I think actually matters, and whether it’s worth your time to upgrade.

TL;DR: This is a meaty bug-fix release. No flashy new features, but it fixes several pain points that have been causing real headaches in production environments — especially around InTouch memory leaks, OPC UA method calls in the Gateway driver, and Historian failover data gaps. If you’re running 2023 R2 SP1 in production, this patch is worth scheduling.

What’s Included

Application Server & OMI: 23.1.102
Communication Drivers Pack: 23.02.102
Historian: 23.1.102
Historian Client: 23.1.102
InTouch HMI (all languages): 23.1.102
OMI Web: 1.0.300.0
Platform Common Services: 8.1.25183.6 (PCS 8.1.5)

Prerequisite: P02 can only be installed on nodes already running System Platform 2023 R2 SP1 or 2023 R2 SP1 Patch 01. You cannot jump straight from 2023 or 2023 R2.

What’s New (Briefly)

Honestly, the feature list is short. The headline item is:

Enhanced OMI Kiosk Mode Protection

Previously, when you enabled kiosk mode on a layout using TitleBarControl, the minimize/maximize/close buttons could still appear if the developer forgot to manually disable them. Now they’re automatically hidden when kiosk mode is on. It’s the kind of fix that should have been there from the start, but better late than never.

Beyond that, you get the usual bundle of hotfix rollups, miscellaneous bug fixes, and some OMI Web client improvements. No new drivers, no new controls, no major architectural changes.

Application Server Fixes (14 Issues)

My pick for the most impactful fix: The multi-developer IDE crash (IMS 3275787). If you’ve ever had your IDE blow up because another user was importing objects at the same time, you know how frustrating this was. It’s finally addressed.

IMS 3275787 — In multi-developer setups with simultaneous activities, aaIDE.exe hung and then crashed, closing the application for all users.
IMS 3306897 — GRAccess-based applications couldn’t be used because template option selections were wrongly reset to defaults, forcing re-embedding of graphic symbols.
IMS 3321814 — No exclamation point icon visible on instances that had errors/warnings.
IMS 3343835 — “Create Instance” for nested templates generated different generic names than previous versions.
IMS 3356780 — Redundant app engine assets appeared twice in the navigation bar; faceplate didn’t display when selected.
IMS 3383589 — After deployment, Engine.Objects dimension and array were not updated.
IMS 3386280 — MQTT Store and Forward data went missing after reconnection.
IMS 3397682 — “Select Alternate Symbol” on child graphics didn’t set the owning object name correctly.
IMS 3416027 — Application Server 2023 got stuck at 81% during check-in.
IMS 3421862 — IO attributes weren’t getting data from the Telemetry Server via PCS.
IMS 3423855 — Objects derived from wizards lost inherited script content after importing unrelated objects.
IMS 3447121 — New internal alarm primitive names were getting reset to old names and settings.
IMS 3474020 — After validating objects on a 2023 P03 galaxy, wizard-derived instances reverted to parent template defaults.
IMS 3447542 — Null reference exception when OPCUA Service clients browsed data.

My take: The wizard-derived object issues (IMS 3423855 and IMS 3474020) are a recurring theme. If you’re heavily using object wizards in your templates, these fixes alone justify the upgrade. The MQTT Store and Forward data loss (IMS 3386280) is also a big one for anyone relying on MQTT for edge-to-enterprise data flow.

AVEVA OMI Fixes (6 Issues)

IMS 3266509 — Graphics and associated text intermittently displayed incorrectly on initial loading.
IMS 3299530 — When using SSRS, the web browser in OMI flipped back to the original date setting instead of using the updated one.
IMS 3425706 — After importing a translation file, OMI didn’t start properly.
IMS 3476034 — Model Drive MES process selection was unavailable when it should have been.
IMS 3479556 — OMI application couldn’t start properly in read-only mode.
IMS 3289435 — The “HideSelf” button on popup windows instantly reopened popups the user had just closed.

My take: The read-only mode fix (IMS 3479556) is important for anyone deploying OMI in viewer-only roles — think control room displays where operators shouldn’t be able to modify anything. The HideSelf bug (IMS 3289435) was one of those “how did this ship?” moments. Glad it’s gone.

InTouch HMI Fixes (30+ Issues)

This is where the bulk of the fixes landed. InTouch got the most attention in this patch, and for good reason — several of these bugs have been causing real operational pain.

The Headline Fixes

Memory leaks are finally addressed. Two separate issues (3804103 and 3636658) caused View.exe memory to balloon — one from 178 MB to 2.4 GB in managed apps, the other in standalone InTouch growing to 2 GB until the app became unusable. If you’ve been scheduling daily View.exe restarts as a band-aid, this patch might let you stop.

Web Browser widget performance (3790790). When multiple Web Browser widgets were used in a graphic, everything shared a single CPU core. This made the entire runtime nearly unresponsive. Huge fix for anyone building dashboards with embedded web content.

All InTouch Fixes

3292661 — Post-upgrade from SP2023 P02 to P04, Industrial Graphics showed AVEVA Logger errors.
3592731 — Japanese keyboard input was garbled in the Japanese version of InTouch HMI 2023 R2 P01.
3636658 — Standalone InTouch: view.exe memory grew to 2 GB, then the app became unusable.
3678282 — Post-upgrade issues with Industrial Graphics and AVEVA Logger.
3692549 — ArchestrAGraphic log messages now include the full graphic name in warnings.
3703479 — Popup last-position not preserved when custom property value pair method was used in scripts.
3704024 — Alarm printer included date/time even when those fields were unchecked in config.
3707212 — Keyboard navigation between screens, quick functions, and scripts was broken.
3715993 — Repeatedly opening/closing windows with Alarm Client Control caused gradual memory increase.
3726241 — AlarmDBLogger connected as Windows user even when configured for SQL authentication.
3730802 — Show Symbol position displayed incorrectly when WindowViewer started maximized.
3732773 — New AD users weren’t propagated to IDE after logging into InTouch Web Client with OS group-based security.
3736501 — Some Galaxy graphics exported as 0 KB files; saving them also errored.
3736915 — Post-migration from 2017 U3 SP1 to 2020 R2 SP1, combo box wrote wrong values to integer tags.
3748177 — Animation switching didn’t work when I/O data came from a Galaxy connection.
3757885 — Post-migration from English to German InTouch, old English key scripts persisted and ActiveX scripts duplicated.
3768399 — “Open windows” button greyed out in WindowMaker when only a few windows were selected.
3770993 — Tag values didn’t update in Industrial Graphic popups.
3771609 — Alarm records disappeared when resizing the graphic window vertically.
3790790 — Multiple Web Browser widgets shared a single CPU core, causing severe slowdowns.
3801478 — Priority couldn’t be added to default query when using Alarm Backup Pair name.
3804103 — Managed InTouch: View.exe memory surged from 178 MB to 2.4 GB.
3811342 — InTouch Web Client symbol placed via iframe into another Web Client didn’t respond.
3827201 — Trend Pen generated excessive logs when querying event history.
3827398 — DBDump failed with 112-character paths in managed apps after upgrade from SP2014.
3834470 — Industrial Graphics with UDTs took extremely long to save after upgrade to SP2023 R2 SP1.
3834546 — ESC key script in FrameWindow threw exception error and didn’t trigger.
3839614 — InTouchDataService.exe NullReference crash took down all Web Client data.
3842871 — Windows 11: symbols with Web Browser widget showed a big red cross and “empty” text.
3853090 — Migration from SP2017 U2 to SP2023 P04: “Unable to generate snapshot” errors in OCMC for all controls.
3873424 — Second domain not shown in View App login dropdown after adding cross-domain config.
3878728 — UDT instance AccessName property overrides not preserved during import.
3895034 — WindowViewer used in-memory cache settings unexpectedly, causing ShowGraphic issues.
3922156 — CALWrapper memory leak: after brief network outage, disconnected topics never came back online.
3924887 — Bitmap import from long file paths showed “Unsupported image file type” error.

Additional fix from System Platform level: IMS-3852895 — Animation value didn’t refresh correctly when a constant expression was used in a FrameWindow symbol script.

My take: The CALWrapper memory leak (3922156) is a sneaky one. It doesn’t show up in normal operation — it only bites you after a network hiccup, which is exactly when you need things to recover gracefully. The fact that disconnected topics just… stayed dead… is the kind of thing that erodes trust in the system over time. Good to see it fixed.

Historian Fixes (11 Issues)

IMS-3842632 — Changing a tag’s description didn’t replicate to PI or CONNECT data services.
IMS-3770292 — Alarm page in Historian Client Web was blank until the scrollbar was moved.
IMS-3800854 — NT SERVICE\InSQLConfiguration wasn’t added to registry during install, causing errors.
IMS-3769882 — aahOleDB crashed in Historian 2023 R2 P01 and 2023 R2 SP1.
IMS-3851701 — Saving another user’s dashboard in Web Client Modern changed its ownership.
IMS-3875123 — Migration from 2017 U2 to 2023 R2 SP1 failed with Int16 overflow error.
IMS-3824648 — Intermittent failover in warm redundancy went into Store Forward, causing data gaps.
IMS-3769395 — Values didn’t populate correctly when loading a new page/window with Historical Playback in OMI.
IMS-3926245 — Unable to migrate Runtime from v11.0 to 2023 R2 SP1.
IMS-3926900 — Silent install showed wrong port (32565) in Configurator even when SilentTCPPort was set to 32560.
IMS-3848494 — Store Forward replication data didn’t transfer to Tier 2 Historian.

My take: The warm redundancy failover issue (IMS-3824648) is the one I’d flag for any operations team. Data gaps during failover defeat the entire purpose of having redundancy. If you’re running warm redundancy configurations, this fix should be your priority.

The replication fixes (IMS-3842632 and IMS-3848494) are also important for anyone running multi-tier Historian architectures. Tag description changes silently not replicating is the kind of bug that causes confusion weeks after the fact when someone notices the downstream data doesn’t match.

Historian Client Fixes (10 Issues)

IMS-3794126 — Some Query tool results could be retrieved in narrow format but not wide format.
IMS-3699452 — Warning message not logged after tag count exceeded 30 in Trend control.
IMS-3855690 — “Value at X2” column values changed length when cursor was at the right limit.
IMS-3832237 — In a dual-VM Historian setup, Client couldn’t failover to the other VM when one shut down.
IMS-3924562 — Alarms remained visible on disabled tags in stacked/single mode layout.
IMS-3902198 — Window Maker showed license check error after migration from InTouch 2020 R2.
IMS-3907815 — Trend control’s BottomY and Top7 properties were invalid in 2023 R2 P01+.
IMS-3838376 — default.aa.Trend file didn’t load properly in embedded Historian Client Trend within managed InTouch apps.
IMS-3557510 — Couldn’t disable file access from Historian Client.

My take: The VM failover fix (IMS-3832237) pairs nicely with the Historian server-side failover fix mentioned above. Together they address a complete failover story for virtualized environments. If you’re running Historian in VMs (and most people are), both fixes matter.

Communication Drivers Pack Fixes

Drivers Core (5 Issues)

3805118 — MMC repeatedly crashed.
3723683 — View.exe hung repeatedly after migration to CDP 2023 R2 SP1 due to a deadlock.
3864132 — Gateway OPC UA datetime values didn’t include milliseconds.
3942846 — Communication Drivers crashed when converting NaN values to Float.
(Internal) — Fixed upgrade issue where driver configuration set was reset to defaults.

Gateway Driver (8 Issues)

3630017 — A single invalid item in a read transaction caused all valid items to get bad quality.
3683666 — [Enhancement] Read request wasn’t sent immediately after a write, causing slower performance.
3814765 — Gateway driver crashed frequently after upgrade due to item count handling.
3870511 — First item (client handle 0x00) wasn’t updated when adding items in OPC Client.
3804935 / 3844400 — [OPC UA] Array input arguments were parsed as strings during method calls, crashing the driver.
3907407 — [Enhancement] MQTT Retain Flag option added to Gateway driver configuration.
3888724 — [OPC UA] Couldn’t read values from OPC UA server when method call returned uncertain values.
3889868 — [OPC UA] Method call wasn’t sent if array argument had no dimensions.

MBTCP Driver (1 Issue)

3800677 — “Swap Byte” option in MBTCP config didn’t work as expected.

SIDIRECT Driver (1 Issue)

3786398 — Default Poke Mode settings didn’t affect Legacy connections.

My take: The Gateway driver got the most love here, and it’s well-deserved. The OPC UA method call crashes (3804935/3844400) were a showstopper for anyone integrating with OPC UA servers that use method calls with array parameters. The “one bad item poisons the whole transaction” issue (3630017) is also a classic — it’s the kind of thing that makes you question your entire data architecture until you trace it back to a single bad tag.

The new MQTT Retain Flag (3907407) is a nice quality-of-life addition for MQTT integrations. It’s a small thing, but if you need retained messages for late-joining subscribers, you previously had to work around it.

The internal fix for configuration sets being reset on upgrade is also worth noting. I’ve seen upgrades go sideways because of this, and it’s not always obvious that the driver config silently reverted.

Known Issues to Watch Out For

Licensing Gotchas

IMS 3978298 — If your node has AVEVA Work Tasks 2023 R2, applying P02 causes a licensing error when launching the IDE. Workaround: Copy Microsoft.Extensions.Logging.Abstractions.dll and PCS.Client.DataStore.dll from ...\AuthenticationProvider\ to ...\Bin\, then restart.

IMS 4014449 — After P02 install, repairing License Server may break AVEVA Connect configuration. Workaround: Delete 4 specific DLLs from the License Server folder, then reconfigure AVEVA Connect.

My take: The licensing issue with Work Tasks (IMS 3978298) is annoying but straightforward to fix. Just be aware of it before you panic when the IDE won’t start after patching. I’d recommend testing the patch on a non-production node first if you have Work Tasks installed.

Historian Known Issues

“Add External Groups” button greyed out in Configurator after install — just relaunch the Configurator
Database Export/Import Utility errors when importing from SP2023 R2 or earlier with saved content/dashboards — upgrade the source Historian to P01+ first
AutoSummary replication Store and Forward path can’t be changed via Configurator — requires manual SQL update to _AutoReplicationServer table
Workgroup environments: secure Client-to-Server communication may fail with DNS enabled — add a hosts file entry mapping the Server IP to its machine name
OData V1 deprecation warning: start migrating to OData V2 now

InTouch HMI Known Issues

IMS-3566480 — Alarm Client Control fails with huge alarm histories — configure MaxTotalRetrievalCount or MaxHistAlarmsCount to limit
4022193 — Trend Pen won’t connect to Historian when SMS is off and InTouch runs on a different machine

Should You Upgrade? My Recommendation

Yes, upgrade if:

You’re experiencing View.exe memory growth in InTouch (managed or standalone) — this patch fixes multiple memory leak paths
You use the Gateway driver with OPC UA method calls — the array parameter crash is a hard blocker
You’re running Historian warm redundancy and have seen data gaps during failover
You rely on MQTT Store and Forward — the data loss on reconnect fix is critical
Your developers work concurrently in the same Galaxy — the IDE crash fix is long overdue
You’re using multiple Web Browser widgets in InTouch graphics — the single-core bottleneck was brutal

You can probably wait if:

You’re on 2023 R2 SP1 P01 and none of the fixed issues affect your environment
You’re still on 2023 or 2023 R2 — you’ll need to get to SP1 first anyway
Your system is stable and you’re in a tight change-freeze window

Upgrade tips:

Back up your Galaxy and Historian databases before patching
If you have AVEVA Work Tasks, prepare the DLL copy workaround in advance
Test on a development or staging node first
Plan for a maintenance window — this isn’t a zero-downtime patch
After upgrading Historian, relaunch the Configurator to avoid the “Add External Groups” greyed-out issue

Final Thoughts

Patch 02 for 2023 R2 SP1 feels like AVEVA is doing the unglamorous but necessary work of shoring up stability. There’s nothing flashy here — no new protocols, no new visualization paradigms. What there is, is a long list of fixes for bugs that real users have clearly been hitting in production.

The InTouch memory leak fixes alone make this worth deploying. The Gateway OPC UA fixes are significant for anyone doing serious OPC UA integration. And the Historian failover data gap fix should give operations teams more confidence in their redundancy setups.

If I had one wish, it would be for AVEVA to bundle release notes in a more digestible format — digging through nested HTML ReadMe files for each component isn’t anyone’s idea of a good time. But that’s a complaint for another day.

Bottom line: schedule the upgrade, test it in your environment, and enjoy the slightly more stable platform.