Have you ever seen Rulerunner skipping all Page nodes producing no error?

In our project, this has been happening once in a while.

In our app, normally, Rulerunner processes  Batch open -> Page open -> Page close -> Batch close.

Once in a while, the same app goes Batch open -> Batch close.

There is no error, just all page nodes skipped.

If I execute the same PDF, this issue won't happen.

My hypothesis is this happens when Rulerunner process cannot use enough CPU or Memory.

------------------------------
dsakai
------------------------------

The RRProcessor intermittently skips Page Nodes and their associated rules. This issue has been reproduced in version 9.1.9 IF006 in my environment using the client app.

I have contacted Support regarding this.

I found an identical case reported in the case database. It was closed, apparently because the frontline engineers were unable to collect the necessary logs.

I was able to reproduce the issue in my environment.

Although this product behavior is rare, I believe it can occur more easily than expected. I'm still curious whether other engineers have encountered this issue before.

Have you ever seen Rulerunner skipping all Page nodes producing no error?

In our project, this has been happening once in a while.

In our app, normally, Rulerunner processes  Batch open -> Page open -> Page close -> Batch close.

Once in a while, the same app goes Batch open -> Batch close.

There is no error, just all page nodes skipped.

If I execute the same PDF, this issue won't happen.

My hypothesis is this happens when Rulerunner process cannot use enough CPU or Memory.

------------------------------
dsakai
------------------------------

Hello Dsakai, 

If a task is being skipped and you can reproduced the issue.  Why not put PageID in it's own thread?  Take a look to see what is happen with that thread.  Review the pageID rrs log and see if there was any issue running the rule.  My guess is most likely it ran out of memory and the batch aborted.  If you are calling any of the recognition rule in PageID, remember to call the release engine.
If the image header X and Y is tool large, you need to resize and rescale the image.  Find out what is causing pageID to not process correctly or why it's taking resources.

You can also look into restart after execute.  After you run pageID you can force the thread to restart which will restarting the RRProcess.exe to clear memory.

Ref: 
https://www.ibm.com/docs/en/datacap/9.1.9?topic=service-configuring-rulerunner-restart-after-task-execution

Seeing restart after execute will take a little longer to process batches.  10 seconds is the default interval for the thread to stop and restart after it process a task.   You can control this in the registry and set it to 5 seconds but I wouldn't go lower then 5 seconds.  Instead of controlling it in rulerunner using restart after execute process.  Look into why it's consuming that much more research.  Again it could be your image.  You'll need to fine tune the tiff or examine are you putting too big of a PDF for convert to process.  Are the PDF assembled with mixed images as in huge color page or header size have large X and Y page size.

------------------------------
Duke Lam

Original Message:
Sent: Tue May 13, 2025 12:24 AM
From: dsakai
Subject: Sometimes Rulerunner skips all Page nodes. It only processes Batch node.

The RRProcessor intermittently skips Page Nodes and their associated rules. This issue has been reproduced in version 9.1.9 IF006 in my environment using the client app.

I have contacted Support regarding this.

I found an identical case reported in the case database. It was closed, apparently because the frontline engineers were unable to collect the necessary logs.

I was able to reproduce the issue in my environment.

Although this product behavior is rare, I believe it can occur more easily than expected. I'm still curious whether other engineers have encountered this issue before.

------------------------------
dsakai

Original Message:
Sent: Thu March 27, 2025 02:59 AM
From: dsakai
Subject: Sometimes Rulerunner skips all Page nodes. It only processes Batch node.

Have you ever seen Rulerunner skipping all Page nodes producing no error?

In our project, this has been happening once in a while.

In our app, normally, Rulerunner processes  Batch open -> Page open -> Page close -> Batch close.

Once in a while, the same app goes Batch open -> Batch close.

There is no error, just all page nodes skipped.

If I execute the same PDF, this issue won't happen.

My hypothesis is this happens when Rulerunner process cannot use enough CPU or Memory.

------------------------------
dsakai
------------------------------

Hello Dsakai, 

If a task is being skipped and you can reproduced the issue.  Why not put PageID in it's own thread?  Take a look to see what is happen with that thread.  Review the pageID rrs log and see if there was any issue running the rule.  My guess is most likely it ran out of memory and the batch aborted.  If you are calling any of the recognition rule in PageID, remember to call the release engine.
If the image header X and Y is tool large, you need to resize and rescale the image.  Find out what is causing pageID to not process correctly or why it's taking resources.

You can also look into restart after execute.  After you run pageID you can force the thread to restart which will restarting the RRProcess.exe to clear memory.

Ref: 
https://www.ibm.com/docs/en/datacap/9.1.9?topic=service-configuring-rulerunner-restart-after-task-execution

Seeing restart after execute will take a little longer to process batches.  10 seconds is the default interval for the thread to stop and restart after it process a task.   You can control this in the registry and set it to 5 seconds but I wouldn't go lower then 5 seconds.  Instead of controlling it in rulerunner using restart after execute process.  Look into why it's consuming that much more research.  Again it could be your image.  You'll need to fine tune the tiff or examine are you putting too big of a PDF for convert to process.  Are the PDF assembled with mixed images as in huge color page or header size have large X and Y page size.

The RRProcessor intermittently skips Page Nodes and their associated rules. This issue has been reproduced in version 9.1.9 IF006 in my environment using the client app.

I have contacted Support regarding this.

I found an identical case reported in the case database. It was closed, apparently because the frontline engineers were unable to collect the necessary logs.

I was able to reproduce the issue in my environment.

Although this product behavior is rare, I believe it can occur more easily than expected. I'm still curious whether other engineers have encountered this issue before.

Have you ever seen Rulerunner skipping all Page nodes producing no error?

In our project, this has been happening once in a while.

In our app, normally, Rulerunner processes  Batch open -> Page open -> Page close -> Batch close.

Once in a while, the same app goes Batch open -> Batch close.

There is no error, just all page nodes skipped.

If I execute the same PDF, this issue won't happen.

My hypothesis is this happens when Rulerunner process cannot use enough CPU or Memory.

------------------------------
dsakai
------------------------------

Do you normally place CustomAction.dll in C:\Datacap\<OCRApp>\dco_<OCRApp>\rules,
or do you put it in C:\Datacap\RRS?

I'm wondering if RRProcessor loads the DLL differently depending on its location,
and whether placing it in a particular folder could cause issues like hanging or failures when executing Custom Actions.

Hello Dsakai, 

If a task is being skipped and you can reproduced the issue.  Why not put PageID in it's own thread?  Take a look to see what is happen with that thread.  Review the pageID rrs log and see if there was any issue running the rule.  My guess is most likely it ran out of memory and the batch aborted.  If you are calling any of the recognition rule in PageID, remember to call the release engine.
If the image header X and Y is tool large, you need to resize and rescale the image.  Find out what is causing pageID to not process correctly or why it's taking resources.

You can also look into restart after execute.  After you run pageID you can force the thread to restart which will restarting the RRProcess.exe to clear memory.

Ref: 
https://www.ibm.com/docs/en/datacap/9.1.9?topic=service-configuring-rulerunner-restart-after-task-execution

Seeing restart after execute will take a little longer to process batches.  10 seconds is the default interval for the thread to stop and restart after it process a task.   You can control this in the registry and set it to 5 seconds but I wouldn't go lower then 5 seconds.  Instead of controlling it in rulerunner using restart after execute process.  Look into why it's consuming that much more research.  Again it could be your image.  You'll need to fine tune the tiff or examine are you putting too big of a PDF for convert to process.  Are the PDF assembled with mixed images as in huge color page or header size have large X and Y page size.

The RRProcessor intermittently skips Page Nodes and their associated rules. This issue has been reproduced in version 9.1.9 IF006 in my environment using the client app.

I have contacted Support regarding this.

I found an identical case reported in the case database. It was closed, apparently because the frontline engineers were unable to collect the necessary logs.

I was able to reproduce the issue in my environment.

Although this product behavior is rare, I believe it can occur more easily than expected. I'm still curious whether other engineers have encountered this issue before.

Have you ever seen Rulerunner skipping all Page nodes producing no error?

In our project, this has been happening once in a while.

In our app, normally, Rulerunner processes  Batch open -> Page open -> Page close -> Batch close.

Once in a while, the same app goes Batch open -> Batch close.

There is no error, just all page nodes skipped.

If I execute the same PDF, this issue won't happen.

My hypothesis is this happens when Rulerunner process cannot use enough CPU or Memory.

------------------------------
dsakai
------------------------------

According to the ddk on custom action pdf, it goes in the RRX.   See if that still fail.  Else you can put it in both location to see if that changes things.  Depending on the result you need to make a note if you ever have to move the apps somewhere else.

IBM Datacap provides two different types of action templates. The primary differences are:
• One action template has a separate DLL and RRX file. The code for the DLL is C# and the RRX contains XML that defines the public actions and contains the help for the actions. The DLL must be registered as described in the development guide. The RRX must exist in the RRS directory within the Datacap install.
• One action template builds the RRX file into the DLL. The code for the DLL is C# and the RRX contains XML that defines the public actions and contains the help for the actions. The DLL is not registered and must exist in the RRS directory within the Datacap install.
When a new version of IBM Datacap is released, any custom actions created for versions of IBM Datacap may need to be updated. These updates typically are to update the .NET version to the version currently used by the release of Datacap and dependencies may need to be re-added to a project for it to compile successfully. Follow the instructions in the download package to migrate your existing actions or use the provided template to create a new action library.
When updating the installed version of Datacap to a new release, it may or may not be required to rebuild the action library. Typically, an action library can be compatible with multiple releases of Datacap but this is not guaranteed. It may be required to rebuild an action to use a newer version of .NET or to access a shared Datacap core component that has been updated in the new release.
Once placed into a Datacap install, the actions created for a custom action library will be visible within Datacap Studio so they can be used within an application the same way as the predefined actions that are provided with Datacap.
An action library must follow standard coding rules such as freeing resources that are no longer required.

Do you normally place CustomAction.dll in C:\Datacap\<OCRApp>\dco_<OCRApp>\rules,
or do you put it in C:\Datacap\RRS?

I'm wondering if RRProcessor loads the DLL differently depending on its location,
and whether placing it in a particular folder could cause issues like hanging or failures when executing Custom Actions.

Hello Dsakai, 

If a task is being skipped and you can reproduced the issue.  Why not put PageID in it's own thread?  Take a look to see what is happen with that thread.  Review the pageID rrs log and see if there was any issue running the rule.  My guess is most likely it ran out of memory and the batch aborted.  If you are calling any of the recognition rule in PageID, remember to call the release engine.
If the image header X and Y is tool large, you need to resize and rescale the image.  Find out what is causing pageID to not process correctly or why it's taking resources.

You can also look into restart after execute.  After you run pageID you can force the thread to restart which will restarting the RRProcess.exe to clear memory.

Ref: 
https://www.ibm.com/docs/en/datacap/9.1.9?topic=service-configuring-rulerunner-restart-after-task-execution

Seeing restart after execute will take a little longer to process batches.  10 seconds is the default interval for the thread to stop and restart after it process a task.   You can control this in the registry and set it to 5 seconds but I wouldn't go lower then 5 seconds.  Instead of controlling it in rulerunner using restart after execute process.  Look into why it's consuming that much more research.  Again it could be your image.  You'll need to fine tune the tiff or examine are you putting too big of a PDF for convert to process.  Are the PDF assembled with mixed images as in huge color page or header size have large X and Y page size.

The RRProcessor intermittently skips Page Nodes and their associated rules. This issue has been reproduced in version 9.1.9 IF006 in my environment using the client app.

I have contacted Support regarding this.

I found an identical case reported in the case database. It was closed, apparently because the frontline engineers were unable to collect the necessary logs.

I was able to reproduce the issue in my environment.

Although this product behavior is rare, I believe it can occur more easily than expected. I'm still curious whether other engineers have encountered this issue before.

Have you ever seen Rulerunner skipping all Page nodes producing no error?

In our project, this has been happening once in a while.

In our app, normally, Rulerunner processes  Batch open -> Page open -> Page close -> Batch close.

Once in a while, the same app goes Batch open -> Batch close.

There is no error, just all page nodes skipped.

If I execute the same PDF, this issue won't happen.

My hypothesis is this happens when Rulerunner process cannot use enough CPU or Memory.

------------------------------
dsakai
------------------------------

According to the ddk on custom action pdf, it goes in the RRX.   See if that still fail.  Else you can put it in both location to see if that changes things.  Depending on the result you need to make a note if you ever have to move the apps somewhere else.

IBM Datacap provides two different types of action templates. The primary differences are:
• One action template has a separate DLL and RRX file. The code for the DLL is C# and the RRX contains XML that defines the public actions and contains the help for the actions. The DLL must be registered as described in the development guide. The RRX must exist in the RRS directory within the Datacap install.
• One action template builds the RRX file into the DLL. The code for the DLL is C# and the RRX contains XML that defines the public actions and contains the help for the actions. The DLL is not registered and must exist in the RRS directory within the Datacap install.
When a new version of IBM Datacap is released, any custom actions created for versions of IBM Datacap may need to be updated. These updates typically are to update the .NET version to the version currently used by the release of Datacap and dependencies may need to be re-added to a project for it to compile successfully. Follow the instructions in the download package to migrate your existing actions or use the provided template to create a new action library.
When updating the installed version of Datacap to a new release, it may or may not be required to rebuild the action library. Typically, an action library can be compatible with multiple releases of Datacap but this is not guaranteed. It may be required to rebuild an action to use a newer version of .NET or to access a shared Datacap core component that has been updated in the new release.
Once placed into a Datacap install, the actions created for a custom action library will be visible within Datacap Studio so they can be used within an application the same way as the predefined actions that are provided with Datacap.
An action library must follow standard coding rules such as freeing resources that are no longer required.

Do you normally place CustomAction.dll in C:\Datacap\<OCRApp>\dco_<OCRApp>\rules,
or do you put it in C:\Datacap\RRS?

I'm wondering if RRProcessor loads the DLL differently depending on its location,
and whether placing it in a particular folder could cause issues like hanging or failures when executing Custom Actions.

Hello Dsakai, 

If a task is being skipped and you can reproduced the issue.  Why not put PageID in it's own thread?  Take a look to see what is happen with that thread.  Review the pageID rrs log and see if there was any issue running the rule.  My guess is most likely it ran out of memory and the batch aborted.  If you are calling any of the recognition rule in PageID, remember to call the release engine.
If the image header X and Y is tool large, you need to resize and rescale the image.  Find out what is causing pageID to not process correctly or why it's taking resources.

You can also look into restart after execute.  After you run pageID you can force the thread to restart which will restarting the RRProcess.exe to clear memory.

Ref: 
https://www.ibm.com/docs/en/datacap/9.1.9?topic=service-configuring-rulerunner-restart-after-task-execution

Seeing restart after execute will take a little longer to process batches.  10 seconds is the default interval for the thread to stop and restart after it process a task.   You can control this in the registry and set it to 5 seconds but I wouldn't go lower then 5 seconds.  Instead of controlling it in rulerunner using restart after execute process.  Look into why it's consuming that much more research.  Again it could be your image.  You'll need to fine tune the tiff or examine are you putting too big of a PDF for convert to process.  Are the PDF assembled with mixed images as in huge color page or header size have large X and Y page size.

The RRProcessor intermittently skips Page Nodes and their associated rules. This issue has been reproduced in version 9.1.9 IF006 in my environment using the client app.

I have contacted Support regarding this.

I found an identical case reported in the case database. It was closed, apparently because the frontline engineers were unable to collect the necessary logs.

I was able to reproduce the issue in my environment.

Although this product behavior is rare, I believe it can occur more easily than expected. I'm still curious whether other engineers have encountered this issue before.

Have you ever seen Rulerunner skipping all Page nodes producing no error?

In our project, this has been happening once in a while.

In our app, normally, Rulerunner processes  Batch open -> Page open -> Page close -> Batch close.

Once in a while, the same app goes Batch open -> Batch close.

There is no error, just all page nodes skipped.

If I execute the same PDF, this issue won't happen.

My hypothesis is this happens when Rulerunner process cannot use enough CPU or Memory.

------------------------------
dsakai
------------------------------

I moved CustomAction.dll to C:\Datacap\RRS and started RRProcessor, but the Page Skip issue still occurred.
Changing the location made no difference.

According to the ddk on custom action pdf, it goes in the RRX.   See if that still fail.  Else you can put it in both location to see if that changes things.  Depending on the result you need to make a note if you ever have to move the apps somewhere else.

IBM Datacap provides two different types of action templates. The primary differences are:
• One action template has a separate DLL and RRX file. The code for the DLL is C# and the RRX contains XML that defines the public actions and contains the help for the actions. The DLL must be registered as described in the development guide. The RRX must exist in the RRS directory within the Datacap install.
• One action template builds the RRX file into the DLL. The code for the DLL is C# and the RRX contains XML that defines the public actions and contains the help for the actions. The DLL is not registered and must exist in the RRS directory within the Datacap install.
When a new version of IBM Datacap is released, any custom actions created for versions of IBM Datacap may need to be updated. These updates typically are to update the .NET version to the version currently used by the release of Datacap and dependencies may need to be re-added to a project for it to compile successfully. Follow the instructions in the download package to migrate your existing actions or use the provided template to create a new action library.
When updating the installed version of Datacap to a new release, it may or may not be required to rebuild the action library. Typically, an action library can be compatible with multiple releases of Datacap but this is not guaranteed. It may be required to rebuild an action to use a newer version of .NET or to access a shared Datacap core component that has been updated in the new release.
Once placed into a Datacap install, the actions created for a custom action library will be visible within Datacap Studio so they can be used within an application the same way as the predefined actions that are provided with Datacap.
An action library must follow standard coding rules such as freeing resources that are no longer required.

Do you normally place CustomAction.dll in C:\Datacap\<OCRApp>\dco_<OCRApp>\rules,
or do you put it in C:\Datacap\RRS?

I'm wondering if RRProcessor loads the DLL differently depending on its location,
and whether placing it in a particular folder could cause issues like hanging or failures when executing Custom Actions.

Hello Dsakai, 

If a task is being skipped and you can reproduced the issue.  Why not put PageID in it's own thread?  Take a look to see what is happen with that thread.  Review the pageID rrs log and see if there was any issue running the rule.  My guess is most likely it ran out of memory and the batch aborted.  If you are calling any of the recognition rule in PageID, remember to call the release engine.
If the image header X and Y is tool large, you need to resize and rescale the image.  Find out what is causing pageID to not process correctly or why it's taking resources.

You can also look into restart after execute.  After you run pageID you can force the thread to restart which will restarting the RRProcess.exe to clear memory.

Ref: 
https://www.ibm.com/docs/en/datacap/9.1.9?topic=service-configuring-rulerunner-restart-after-task-execution

Seeing restart after execute will take a little longer to process batches.  10 seconds is the default interval for the thread to stop and restart after it process a task.   You can control this in the registry and set it to 5 seconds but I wouldn't go lower then 5 seconds.  Instead of controlling it in rulerunner using restart after execute process.  Look into why it's consuming that much more research.  Again it could be your image.  You'll need to fine tune the tiff or examine are you putting too big of a PDF for convert to process.  Are the PDF assembled with mixed images as in huge color page or header size have large X and Y page size.

The RRProcessor intermittently skips Page Nodes and their associated rules. This issue has been reproduced in version 9.1.9 IF006 in my environment using the client app.

I have contacted Support regarding this.

I found an identical case reported in the case database. It was closed, apparently because the frontline engineers were unable to collect the necessary logs.

I was able to reproduce the issue in my environment.

Although this product behavior is rare, I believe it can occur more easily than expected. I'm still curious whether other engineers have encountered this issue before.

Have you ever seen Rulerunner skipping all Page nodes producing no error?

In our project, this has been happening once in a while.

In our app, normally, Rulerunner processes  Batch open -> Page open -> Page close -> Batch close.

Once in a while, the same app goes Batch open -> Batch close.

There is no error, just all page nodes skipped.

If I execute the same PDF, this issue won't happen.

My hypothesis is this happens when Rulerunner process cannot use enough CPU or Memory.

------------------------------
dsakai
------------------------------