Datastage PX Config file setup error

sudhirmah · Post by **sudhirmah** » Tue Mar 29, 2005 2:18 am

I have tried to change the the PX configuration file from 2 nodes to 3 nodes, am getting the following error, pls advise,TIA

XFM_Rec_Count: Failed to load the library "V0S117_LOADCLSCMDimStagingRecon_XFM_Rec_Count.o". Either the directory containing the library file is not on the library search path, or the library was compiled on a system that is incompatible with this system. Could not load "V0S117_LOADCLSCMDimStagingRecon_XFM_Rec_Count": 0509-130 Symbol resolution failed for /baselwb/ds_project/HydraQDW/RT_BP294.O/V0S117_LOADCLSCMDimStagingRecon_XFM_Rec_Count.o because:
0509-136 Symbol UvalueStrHandleHex__19APT_EnvironmentFlagCFv (number 166) is not exported from
dependent module /baselwb/dsadmhome/Ascential/DataStage/PXEngine/lib/liborchcoreaix3.a(libtemp.o).
0509-192 Examine .loader section symbols with the
'dump -Tv' command..

ArndW · Post by **ArndW** » Tue Mar 29, 2005 3:11 am

Can you post the new configuration file and tell us which node you added?

sudhirmah · Post by **sudhirmah** » Tue Mar 29, 2005 3:39 am

Below is the config file

=======================================
main_program: APT configuration file: /baselwb/dsadmhome/Ascential/DataStage/Configurations/HYDRA.apt
{
node "node1"
{
fastname "sgwbba05"
pools ""
resource disk "/baselwb/hydra1/Datasets" {pools ""}
resource scratchdisk "/baselwb/hydra1/Scratch" {pools ""}
}
node "node2"
{
fastname "sgwbba05"
pools ""
resource disk "/baselwb/hydra2/Datasets" {pools ""}
resource scratchdisk "/baselwb/hydra2/Scratch" {pools ""}
}
node "node3"
{
fastname "sgwbba01"
pools ""
resource disk "/baselwb/ds_wdir/hydra3/Datasets" {pools ""}
resource scratchdisk "/baselwb/ds_wdir/hydra3/Scratch" {pools ""}
}
}
=======================================

tried to narrow down the problem...its with db2 shared library, following is the error.
APT_CombinedOperatorController,0: Fatal Error: Fatal: Shared library (dsdb2.so) failed to load: errno = (2), system message = (No such file or directory)

roy · Post by **roy** » Tue Mar 29, 2005 4:29 am

Hi,
at first glance it seems that DS can't find the specified directory/ies in the configuration file you mentioned for node3.

IHTH,

T42 · Post by **T42** » Wed Mar 30, 2005 10:39 pm

Be careful with using different names for your fastname. This would automatically cause DataStage EE to think that it's running as a MPP process, and will try to transfer the Transformer libraries over to the sgwbba01 server via RSH (see $APT_ORCHHOME/etc/distribute-component - that's the script that is being used on UNIX). If RSH fails, OR the path is non-existing, or you don't have permission, or the mountpoint doesn't exist... it does.

Now in a MPP environment, the directories on ALL servers must be precisely the same. Also, the PX engine must exist on all servers.

This is true for remote DB2 setups, btw. It is in fact a MPP system, with DB2 server as the 2nd (or whatever) server. This is part of the reason why I suggest that you get the latest copy of the DB2 configuration guide (I forgot the name of that document offhand) from Ascential.

Alrighty, class dismissed. My solution: Fix the fastname to match the other fastnames. I don't think you were going for a MPP system.

goma · Post by **goma** » Thu Mar 31, 2005 3:08 am

I have got similar message after importing job from the dev box to production box. It is caused by different version of C++ compiler between dev and prod box. After compiling job, it become ok.