Please include the complete stack trace. It would be helpful for folks to see all the stack frames. ... View more

If I was in your shoes I'd have a look at the API trace that can be generated with sastrace. You might try use the same ODBC driver in another application (e.g. isql from unixODBC). ... View more

@imdickson I recall seeing a bug ticket about this the other day. It was for a customer in Asia. Having a NaN value in your data causes a known issue in LASR that has been there a long time. It only recently was seen being triggered by a customer. Both cases I've seen now involved data in LASR being loaded from PostgreSQL to LASR and a +infinity was in the table. It is actively being looked at by R&D. ... View more

Your questions are best put to your Microsoft Windows system administrators. The configuration of your Microsoft Windows virtual memory subsystem is best left to your Windows experts. For most (if not all) SAS applications you don't want any paging occurring. There is a good guide by Moti Bani from Microsoft here that I suggest you read.   The Windows page file is used as the dump location for all of memory if the system panics. When you have large amounts of memory this demands a larger page file.   > 1. What is using the C drive if the SAS Application is installed on D, users access it via Citrix and SASWORK points to E (SAN disk)?   The application needs the operating system to interface with the hardware. The C:\ would contain all the binaries and the kernel used by the application to interface with the physical hardware. Your sysadmins have pointed out there are key files used by the operating system like the page file that were configured by your Microsoft Windows system administrators to reside on that volume.   > Why is the application paging if RAM is so high on the server, but heavily unused?   Having a large page file on Windows does not necessarily mean the virtual memory system is paging. The performance monitor exposes the page faults/sec in Windows. That page fault/s is major and minor faults, no idea who to track major faults only. Not sure what the Windows equivalent of vmstat(1) is but I'd be sure Microsoft exposes the paging rate somewhere for users to check.   > Though the recommendation for the Page File is usually 1.5X RAM,   This was a rule of thumb some time back but has not been the case for a while. For me since the swap to 64-bit systems some 10 years ago that 1.5X RAM rule of thumb isn't something I follow.     I don't use Windows much, most of my experience is with Linux, so consider my advice with that in mind. For me personally I typically don't allocate any swap if I know I have enough memory for the application. If the kernel starts to get unstable and I need to capture crash dumps then I'll make a plan to attach some disk. For day to day running swap isn't needed when you have lots of memory. ... View more

Typically the simplest approach to revert a hotfix is to restore the system from a backup to a state before the hot fix was applied taken . With virtual machines I typically take a volume snapshot before I start to apply the hot fix. With systems that have a filesystem that can provide a snapshot (like ZFS) I'll use a filesystem snapshot.   If anything goes wrong its simple to get the machine(s) back to the state they were in before things went sideways. ... View more

Protip - in your bashrc add alias sas-grid='sas-grid-cli.sh --mhost <master-host> --mport <port> --grid-user <username> --grid-password <password>'   Once you've got that done you can hammer out those grid queries really quickly.   sas-grid-cli.sh --mhost <master-host> --mport <port> --grid-user <username> --grid-password <password> show-jobs ... turns into ... sas-grid show-jobs ... View more