mroz@moose.UUCP (09/30/87)
[] I'm having a problem with the use of paging file pages on different machines. On any one of our nonclustered, VMS 4.5 Vaxen it costs about 600-700 pages to issue the DCL SPAWN command. However, on a clustered system (four nodes), spawning a subprocess requires about 1600 pages. To calculate this I did the following: $ sh proc/quo $ spawn $ sh proc/quo and subtracted the second paging file quota from the first one. Are the extra 1000 pages needed because the system is clustered? The clustered and nonclustered systems both set up an equivalent amount of logical names and symbols (I think!). My quotas (open files, subprocesses, etc.) are identical on the different systems. I've talked to our system managers and they're not sure where to look. Can anyone shed light on this? I need to know what sort of system setup parameters are important. To avoid cluttering up usenet please e-mail me directly. Thanks a bunch, Peter Mroz General Electric | ARPA: mroz@ge-crd.arpa Corporate Research and Development | UUCP: mroz@moose.steinmetz.ge.com PO Box 8, 37-2081 | clever saying stuck in here Schenectady, NY 12301 | 518-387-6021