Re: MCH0601 with QMHRDQM in V7R1

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

You can report this to IBM as a bug if you like....  IBM should not have 
broken compatility w/V7R1.

But, I know that they changed the way stuff works with 6.1 to optimize 
many things...  So now that you've made me curious, I took a quick look 
at the docs, and I see the following note:

   Note: If not using the optimum space alignment, the user space
   has a maximum size of 16MB minus 512 bytes (16,776,704 bytes).
   If optimum alignment is specified, the maximum size of the
   user space is 16MB minus one disk page (current page size is 4096
   bytes, giving a maximum space size of 16,773,120 bytes).

This is from the following page:
http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/apis/quscrtus.htm

Elsewhere on that page, you'll see that there is a parameter where you 
can specify whether you want to have (or not have) optimum alignment. It 
is an optional parameter that you are not passing.

According to the docs, if that parameter is not passed on V5R4, the 
default is 0 (non-optimum alignment).  On V6R1, it's 1 (optimum 
alignment).

Obviously, this means that if you do not pass the parameter, it'll run 
faster on 6.1, but the maximum size will be 4096 bytes smaller.

Personally, if I truly need a large size, I'd use teraspace (as I said 
before), this allows for a MUCH larger size, so would be great.

But, if 16773120 will work, you might just use that.  I can't imagine 
the 4k matters too much to you, does it?


On 8/22/2013 4:10 PM, Shubhadeep Roy wrote:
> Hi Scott,
>
> I have never worked with teraspace allocation and honestly do not know how
> to do that. As per the IBM API manual, the Initial Size parameter can have
> a value from 1 to 16776704 and that is what I have used. The purpose of the
> utility is to read through any data queue (keyed/non-keyed) and store the
> contents in a SQL table and also clear the messages from the data queue, if
> the user so chooses. This works absolutely fine in V5R4, I have tested this
> with various queues in different libraries and also on different systems.
>
> If you could point me to an alternative/better way to do this, I would be
> grateful.
>
> Thanks.