Yeah that is what IBM support told me, and they provided me with a snippet to go in paw-override.yml.
I haven't found any other way of fixing it but that is mostly because I have had other things to work on.
Cannot comment on performance but I didn't see a noticeable drop-off after making the change.
Original Message:
Sent: Mon September 15, 2025 07:11 AM
From: Kasper Dueholm
Subject: PAW Action Button Process Aborted but Actually Runs Fine
Hi.
We have seen this issue with a customer too - and disabling HTTP/2 "resolves" the issue.
We noticed that if we use a Chromium based browser from within Azure it will work - issue is only occurring from a client outside Azure.
(Firefox work from both outside and within Azure).
Disabling HTTP/2 is more like a workaround rather than a fix - and our customer is worried about performance.
IBM Support is not recognizing this as a defect, and suggest making a RFE.
Has anyone found a better solution to this problem?
BR,
Kasper
------------------------------
Kasper Dueholm
CogniTech A/S
Original Message:
Sent: Wed September 25, 2024 09:56 AM
From: Larry Ding
Subject: PAW Action Button Process Aborted but Actually Runs Fine
Hi Tomáš,
For PAW local, it can be disabled via a PAW deployment override, with the following steps:
1 . Edit or create paw/config/paw-override.yml
2 . Contents:
3. Apply changes: scripts\paw.ps1
------------------------------
Larry Ding
Original Message:
Sent: Wed September 25, 2024 09:39 AM
From: Tomáš Vojta
Subject: PAW Action Button Process Aborted but Actually Runs Fine
Hi Larry,
we seem to have the same issue, so I'm curious, how to disable http/2 for PAW local. I didn't find any clue in documentation, could you help please?
Thanks a lot
tv
------------------------------
Tomáš Vojta
Original Message:
Sent: Fri July 26, 2024 09:06 AM
From: Larry Ding
Subject: PAW Action Button Process Aborted but Actually Runs Fine
Hi Tom,
Is the Azure VM Windows server 2016? Were you using Chrome/Edge? Does Firefox help?
It could be caused by http2 issue between Chromium based browsers and the library PAW uses on Windows server. Try Chrome with http2 disabled(https://my.f5.com/manage/s/article/K000092500).
Http2 can be disabled from PAW local if it's confirmed to be the cause.
------------------------------
Larry Ding
Original Message:
Sent: Thu July 25, 2024 12:39 PM
From: Tom Cook
Subject: PAW Action Button Process Aborted but Actually Runs Fine
Hi Stuart,
The environment is an Azure VM which is hosting PAW, TM1 and Cognos Analytics.
ProxyTimeoutSeconds is set to 900.
This is a PAW workbook with a PAW action button, no websheets involved.
Thanks,
Tom
------------------------------
Tom Cook
tom.cook@actionkpi.com
Original Message:
Sent: Thu July 25, 2024 08:17 AM
From: STUART KING
Subject: PAW Action Button Process Aborted but Actually Runs Fine
Hi Tom,
The action button in PAW make async REST API calls to TM1, so unless there is a very low timeout set (e.g 10 seconds) this should not happen. Note that the PAW cube viewer and most other interfaces in PAW also make async API calls to TM1 as well.
I might suspect there is some network infrastructure (software or hardware) between PAW and TM1 that is closing the connection.
Also, just to clarify, this is a PAW action button on a PAW book? Not a Websheet button on a Websheet in a PAW book?
------------------------------
Stuart King
IBM Planning Analytics Offering Manager
Original Message:
Sent: Wed July 24, 2024 03:54 PM
From: Tom Cook
Subject: PAW Action Button Process Aborted but Actually Runs Fine
Hey everyone,
Running into an issue where, when I run a process from an action button in PAW 2.0.96 (and other earlier versions), a notification pops up that the process was aborted after 10 seconds, but there are no error logs and the process runs fine.
Funny thing is, if i have another browser window open on the Admin pane, auto-refreshing the server thread monitor every few seconds, or if I have a cube view on the PAW tab where the action button is and set it to refresh every 5 seconds or so, then the process runs fine and notifies, after the ~30 seconds it takes, that it has run fine.
Is there some esoteric timeout setting I am just not configuring that needs to be configured that is responsible for this? Seems like there is an issue with PAW just NOT keeping the connection with TM1 open, or perhaps more likely the browser not keeping the connection with PAW open, for anything longer than 10 seconds in this instance.
Thanks,
------------------------------
Tom Cook
tom.cook@actionkpi.com
------------------------------