1. Where do I find the error code descriptions. I read a note in the online discussion from Aaron that they are listed in the sysinclude file, but… my knowledge level does not direct me just where the susinclude file is. Since we went to NT we get periodic, seemingly out of the blue, fatal errors and I would sure like to stop getting them. We use AREV 2.12.

2. We get (and had in the Novell days) on some select commands duplicate records. Our main customer file, for example. When I run a "Select CM with ….." I can can get the same customer listed 6 times, and another just once. There is only one of each, so what's going on? I have written a piece of code after the the readnext to skip a record if the number is the same, but that just does not seem the proper way, and besides, there are lots of reports.

1. Where do I find the error code descriptions. I read a note in the online discussion from Aaron that they are listed in the sysinclude file, but… my knowledge level does not direct me just where the susinclude file is. Since we went to NT we get periodic, seemingly out of the blue, fatal errors and I would sure like to stop getting them. We use AREV 2.12.

If I understand correctly, what you are looking for is in the REVERROR file, which is a ROS type file located in the directory where your AREV.EXE is located. Normally, the descriptions of these error messages are displayed to the user. However, if the workstation has insufficient RAM, your computer will be unable to allocate memory to display the message. In this case all you get is an error code.

To view the contents of the REVERROR file, create a ROS type volume (all you need to do for this particular case is create a VOC item with VOLUME on the first line and RTP51 on the second line). Now, set an alias ("SETALIAS ROS_FILES SYSPROG,sysprog_password REVMEDIA REVMEDIA"). Then "EDIT REVMEDIA REVERROR*GLOBAL". On the first line, type "REVERROR.0". Save the record. Now you can attach REVERROR on the volume you created and use it as you would any other file.

CAUTION: Don't update this file. ROS files should not be written to in a multiuser environment. They should be considered read only files.

2. We get (and had in the Novell days) on some select commands duplicate records. Our main customer file, for example. When I run a "Select CM with ….." I can can get the same customer listed 6 times, and another just once. There is only one of each, so what's going on? I have written a piece of code after the the readnext to skip a record if the number is the same, but that just does not seem the proper way, and besides, there are lots of reports.

I'll wager you are doing a sort by a multivalued field somewhere in your select. This has the effect of expanding the select to one record per value.

Victor:

Thanks for the response.

Here is more:

1. The most common fatal error we get is :"Error 2102 returned during RCL request". Another one we see is "Error 4020 returned during RCL request". What do they imply? Again these never happenned in Novell, only in the NT environment.

2. Concerning the multiple records, the select statement is very simple: "Select CM with". CM is the customer master file, and uses integer values as the key. ID is also single value. When indexing is ON, the report returns one record per ID; when indexing is OFF, it can return one to six records, depending on what account is chosen. Those accounts that return multiple copies are consistent in doing it, but again, when indexing is on, they return but one.

Thanks,

Jim

1. The most common fatal error we get is :"Error 2102 returned during RCL request". Another one we see is "Error 4020 returned during RCL request". What do they imply? Again these never happenned in Novell, only in the NT environment.

These look like something other than Arev messages. I know next to nothing about NT so will let someone else comment.

2. Concerning the multiple records, the select statement is very simple: "Select CM with". CM is the customer master file, and uses integer values as the key. ID is also single value. When indexing is ON, the report returns one record per ID; when indexing is OFF, it can return one to six records, depending on what account is chosen. Those accounts that return multiple copies are consistent in doing it, but again, when indexing is on, they return but one.

Do you have an MFS on this file? By "account" are you referring to an Arev account (application) in which case you'd be looking at a different file?

These errors are related to the NPP, however I'm not sure exactly what they mean. They seem to indicate a lack of communication with the server, a timeout or invalid route or something like that.

apk@sprezzatura.com

Sprezzatura, Inc.

I solved these error problems. Because often the problem related to an !file or f.distributor, or randomly appeared, I looked at my index settings. The default settings f10-Management-Environment-Indexes was 60 second delay before indexing, and 30 seconds before index checks. I dropped the time back to 180/90, and the problems have disappeared. I gather NT does not handle frequent index checking as well as Novell. Also, odd… when I changed the default to 1 for Time delay between indexes, I got a strange case of sporadic keystroke delay…. in which I could type a word with the screen not responding, and then catching up. Sometimes this would happen, and other times all would function normally.

So I hope this can help other lost souls. Thanks for the input everyone.

Jim

View this thread on the forum...