[WEB400] Re: Locking problems using CGIDEV2

[Mel Rothman <mel@xxxxxxxxxxxxxx>]

I suspect that this is not a CGIDEV2 problem.

More likely, a PTF is needed.  PTF SI13570 for V5R3 5722DG1 addresses CGI job 
threads getting into a deadlock situation.  See:

http://makeashorterlink.com/?D1CB21F29 for details.

If that isn't it, I suggest that you gather detailed information about the 
messages (message ID, sending and receiving programs, modules, functions, 
statement numbers, etc.) and report them to IBM support.

If a CGIDEV2 component is receiving the message, then it's possible there is a 
CGIDEV2 problem.  If so, please make sure you have the latest CGIDEV2 version 
(currently 2003-09-01 14:50:19).

If you discover anything that indicates the problem is in the current CGIDEV2 
version, I will be happy to fix it as quickly as possible.


Mel Rothman, CGIDEV2 Author
Mel Rothman, Inc.

Mike Skvarenina wrote:
> For several years now my shop has been very successfully using Mel's
> CGIDEV2 library to create web enabled apps on our iSeries.  From time to
> time however, at least several times a week, some sort of lock will
> occur and it will bring down the entire HTTP subsystem.  When we work
> with a job that's in a LCKW status, we don't see anything typical like a
> file or record wait.  The ONLY clue we find in the jobs sitting in LCKW
> status is a message in the joblog: "Invalid space location unlock
> request.".
>
>  
>
> Since in a very short amount of time we could have 100 of these HTTP
> jobs sitting in LCKW status, the fastest way for use to resolve it is to
> end the subsystem, wait for all the jobs to end, then restart the
> subsystem.
>
>  
>
>  
>
> Has anyone else seen this problem and if so how'd you solve it?
>
>