Re: [CMS-PIPELINES] FPLRVR236E Too much data for variable MSTEM.5020
On Wed, 24 Aug 2022 at 16:32, Stanislawski, Shawn (National VM Capability) < shaw...@dxc.com> wrote: > 16MB? 24bit addressing related somehow, is it? > But yes, it is very likely the assignment of this REXX stemmed variable > attempts a storage request over 16MB (including control information). I > will check. > Since you have the data in the pipeline to write to a stemmed variable, you could block it conveniently before putting it there, and deblock it when you use it. But it could be very attractive to put the application logic in the pipeline and avoid staging the data in REXX variables. > Just to be thorough, are there limitations regarding the area of CMS > available memory where the REXX variable pool can reside? (16 MB line, 2 > GB bar, etc) > > The total amount of REXX variables (plus what else you have) needs to fit under the bar in your virtual machine. Rob
Re: [CMS-PIPELINES] FPLRVR236E Too much data for variable MSTEM.5020
On 8/24/22 16:32, Stanislawski, Shawn (National VM Capability) wrote: 16MB? 24bit addressing related somehow, is it? Not likely, even though REXX is before XA, but MVCL &c are restricted to 16M objects.
Re: [CMS-PIPELINES] FPLRVR236E Too much data for variable MSTEM.5020
16MB? 24bit addressing related somehow, is it? But yes, it is very likely the assignment of this REXX stemmed variable attempts a storage request over 16MB (including control information). I will check. Just to be thorough, are there limitations regarding the area of CMS available memory where the REXX variable pool can reside? (16 MB line, 2 GB bar, etc) --Shawn S.
Re: [CMS-PIPELINES] FPLRVR236E Too much data for variable MSTEM.5020
A single REXX variable is limited to 16M because of how REXX works. When you're doing a REXX stemmed variable each shorter than that, you can fill up CMS available memory with them. Sir Rob the Plumber On Wed, 24 Aug 2022 at 00:40, Stanislawski, Shawn (National VM Capability) < shaw...@dxc.com> wrote: > Have a situation where we keep running into FPLRVR236E when trying to > populate very large stemmed arrays. > Any idea what maximums the STEM stage / REXX itself is capable of handling? > > (If it helps: virtual machine currently set to MACHINE XA, 1G of virtual > storage, IPL'ed CMS saved segment.) > > > --Shawn S. >
