I sometimes get a load error when my RC jobs run. I am more than willing to log a case on this, but I thought I'd throw it out here first just to see if anyone had any thoughts or had come across anything similar. The odd thing is, it's not consistent -- I can get the error, turn around an run it manually (through RC) and it works fine. At first I thought it was due to the amt of activity going on, so I started spacing my jobs out -- but I think what I saw this morning sort of shot that idea out of the water...

Here's a snippet of what shows up in the RC log:

BTP1020 AppServer command :OLD-STYLE COMMAND: invalid in state :MAINTAIN
BTP1020 CALL SERVER:
BTP1020 UNRECOGNIZED COMMAND CASE TOP;
BTP1020 UNRECOGNIZED COMMAND AMEASURE_SERIES/A11 = 102 ;
BTP1020 UNRECOGNIZED COMMAND CALL MEAS_LOADER FROM AMEASURE_SERIES;
BTP1020 UNRECOGNIZED COMMAND ENDCASE;
BTP1010 Task error: AppServer command :OLD-STYLE COMMAND: invalid in state :MAINTAIN


Again, I'll probably go ahead and log this -- but I thought I'd ask here first just in case.
Thanks!

This usually happens if AGENTREFRESH is set too high... WebFOCUS Maintain sometimes leaves agents in a state where they can't run other maintains. That's an oversimplification, but basically if you set AGENTREFRESH to 1 this problem goes away. If you were a big Maintain shop it would be a bigger issue and more intricate to solve, but since you only are only use WebFOCUS Maintain inside PMF, it's a simple change of setting.

You will find that setting on the Service window in your WebFOCUS Server Console. Typically you find this on your machine/port 8121, but if you're not on Windows, it's usually another port.

Looks like ours is still at the default value (100).

Are there any disadvantages of setting this down to 1? It seems like there might be some efficiencies to be had by not refreshing every agent after one connection -- but then again I also want my loads to work.

Just curious if I should expect any other changes in behavior...

If you had many older persistent Maintain apps onsite, and these apps did specific things, you might see behavior you wouldn't like, so would need to modify other things in your environment.

But that's not the case at your site for this server. PMF has always used Maintain non-persistently. So no ill effects.

There is no real benefit to having agentrefresh be set to 100, except that it saves a few extra milliseconds when agents don't refresh. Refreshing them does ensure a nice clean agent for each request.

In the future this need will be phased out. It's only a temporary situation.

Without knowing what you have in your RC procedure it sounds like you have some MODIFY commands in a maintain procedure perhaps? Maybe in preprocessing procedure? This is what I found in the IBI Tech database http://techsupport.informationbuilders.com/sps/72201091.html
What is MODIFY and where is it documented?
Solution:

MODIFY is the old-style WebFOCUS command for creating/modifying
records in a database. While it has since been replaced by MAINTAIN as
an interactive database maintenance language (DML), it has been retained
in the product for upwards compaatiblity issues.
There is no documentation in the WebFOCUS manuals for a variety of reasons:

1. WebFOCUS is primarly reporting, not database creation;
2. MAINTAIN has replaced MODIFY as the interactive DML of choice;
3. MODIFY can be complex, depending on its usage.

That being said, all the Core FOCUS manuals (IBM, Unix, VMS) still document
the command. These manuals are available from your branch, or may be
downloaded from the IBI web site

thanks Rhonda; that's not the case in this case - Trav is talking about the loader in PMF which is 100% Maintain (for now).

But we appreciate your help.

Bob, do you think that changing this setting could affect the graphing engine in any way?

I'm not totally convinced that it's related, but it seems like when I change that setting, we start to have problems with our graphs that generate SVG hold files. We get some strange JAVA errors (I don't have the exact error right now). It's happened a couple of times now and I've ended up switching that setting back to 100 both times.

Unless you think it IS specifically related, I'm going to switch it to 1 again and see if I can reproduce the problem and then probably log a case.