DSXchange

Ours could be the longest possible upgrage to be done ever. We are upgrading from 6.0.1 to 7.0.1r1 and this is not an inplace upgrade. We installed 7.0.1r1 in our new server and imported all the jobs from 6.0.1 server. Once the import is complete, we are having tonnes of problems in running the jobs successfully. Could anyone tell me the best practice to do this kind of upgrade. The other point to be mentioned here is that we had to move the datasets from 6.0.1 server to 7.0.1r1 server to have our jobs run. We converted the datasets to .dat's and ftp'd them to 7.0.1r1 server and reconverted them to datasets. The config. file is the same for both the servers. The path / locations of all the datasets / files etc remain the same on both the server. The only thing that changed in the config. file is the path to the scratchdisk. Could someone give me the info as far as different settings to be done at the server and project level.

Thanks

Netboyks

Your post contains alot of *missing* information. We can best help you if you can further describe both server environments, database/network connections to<>from both, etc.

Our upgrade was painless and have been testing for about 4 weeks now with no problems whatsoever.

Mine have always been painless as well, using either the upgrade-in-place or the new-install-and-import methodologies. However, these were Server upgrades, not PX, so I can't comment on the ease of those.

As has already been 'pointed out' more information would help, including perhaps some examples of your "tons of problems" you are seeing.

I've just done one from Solaris to AIX, DS 5.2 to 7.0. The main obstacle was changing default values for pathname parameters; which I solved by creating a DataStage routine that post-processes a DSX-format export file.

Thanks for your replies. The situation is that we are upgrading 3 things here. Solaris 2.8 to 2.9, Oracle 8i to 9i and DS 6.0.1 to 7.0.1r1. After installing solaris and oracle we installed DS and imported the jobs from 6.0.1 and compiled them. The values that changed are the path of the DSEngine, the name of the projects. We also changed the path for the projects in Unix by creating a seperate mount point. The other paths for the datasets, files, database did not change. The path for the config. files remain the same. To run our jobs we had move the datasets from 6.0 to 7.0 server. We created a seq. file out of the dataset, ftp'd the file and recreated the dataset using the same config. file. These are the changes we did during the upgrade and we are having problems with most of our jobs running. Could some one let me know if there is something we need to do to fix.

Thanks

Netboyks.

You can't just ftp datasets between machines. Use the dataset management tool to move them (manager -> tools etc ...)