12-04-2014 02:36 PM
I am using SAS VA 6.4 non-distributed version on windows environment. I defined multiple LASR libraries for which I set up separate autoload scripts. It works fine until LASR server is stopped and autoload mechanism have to start it by itself. I receive below error:
ERROR: A connection to the LASR Analytic Server on 'XXXXXXX', port xxxxxx, could
not be made. Make sure that the host and port are correctly specified, that you are attempting
to connect to a LASR Analytic Server of the correct vintage, and that the server is still
ERROR: Libref LASRTEST is not assigned.
ERROR: Error in the LIBNAME statement.
NOTE: LASR server xxxxxx:xxx is not running.
NOTE: Starting single machine signed LASR Server xxxxx:xxxx
NOTE: There are no active tasks/transactions
LASR server is started by autoload process but next autoload processes ended with below error:
ERROR: Insufficient resources to perform the operation.
and following autoload processes not refresh in memory data correctly.
I thought that maybe there in not enough memory when it have to reload data but I did below check:
- I stop my scheduled autoload processes.
- I stop LASR server from va hub
- I start LASR server from va hub
- I schedule my autoload processes
I noticed that loading in to memory finished successfully without any errors. I think there are differences on command parameters which is send when LASR server is started using HUB comparing to autoload process. I found that command from va hub is using two additional parameters: memsize 0 and cpucount actual when comparing what is set on autoload mechanism but I do not think it is source of my problem and that setting those parameters will solve an issue.
Can anybody spot this issue? Maybe any ideas where I can look to solve this problem? R
03-15-2015 07:52 PM
i *do* think that the problem can be caused by not having memsize 0 and cpucount actual when autoload runs.
i ran into the same problem (ERROR: Insufficient resources to perform the operation) and see that the sas autoload session runs with memsize 2G, where the workspace server that is started by the hub has memsize 0 = unlimited memory.
i'm trying to find the best place to set the memsize for all autoload scripts that we run, there doesn't seem to be a placeholder provided.