Too Much Free Memory on a VAX V7.1

[Steve Reece_3]

>>>Steve, they move to the Modified List first, not to the Page File(s).<<<

Nope, still confused. Sorry to be clueless on this.

[Hoff]

>Nope, still confused. Sorry to be clueless on this.

Pages of memory that overflow (page out) from the working set out to the modified list and can move from there out to the pagefile (slow) or back into the working-set.

Faults from the modified list are so-called soft faults and (while there's some overhead here) not nearly as expensive as a so-called hard fault of a memory page back in from disk.

See "7.2.1.1 Hard and Soft Page Faults" here:

http://h71000.www7.hp.com/doc/73final/6491/6491pro_006.html

[Robert Gezelter]

Janet,

Understanding the usual dangers of "prescribing over the phone" with many of the details unknown, I note that rather than increasing the quotas, an often higher-performance pathway is to do the needed homework to INSTALL the images used by the application.

INSTALLED shareable images are far and away a huge performance win if the majority of the pages are image pages (as opposed to read/write data pages).

- Bob Gezelter, http://www.rlgsc.com

[oldskul]

Again, everyone. Thank you.

I am at this very moment creating the presentation notes for a meeting tomorrow to change WSMAX on the cluster. One baby step is better than no baby steps.

I installed every image I could last year though we still have many open files on the system disk, it does not have an IOREQ Len greater than .10 ever.

We have a high image activation and some of those are really large cobol programs that are not written with predication in mind.

The biggest system bovine we have is the 4GL FOCUS from IBI. Application design is not VMS saavy.

The EDA user is a WebFocus (TSCOM3) interface. This is the EDA user, which by the way, are detached processes waiting for web users. I apologize if I repeated this.

Reply only gives me the original question. I'm new. I'll get with the program and figure out a better way to answer.