No Data Selection?

[vasuk]

We have a custom UBE running in scheduler, in Xe. Although it has run many times OK it just appears to have run without its data selection. Well at least it processed some records that would be excluded by the data selection.

Can't help thinking I have heard of this before and the problem is incomplete spec's and the solution is to delete and re-create the version.

Can anybody confirm this and/or explain what causes the spec's to get corrupted?

Thanks,

Mike

 

[Segfault]

Most likely, someone ran the same version from a client box with open data selection. This changes the server's specs for the version, and the next time scheduler runs it, it will run with the changes.

Scheduler versions ned to never be run by anyone but the scheduler once the DS / POs / Sequencing are properly set up, or stuff like this can happen.

 

[DSauve]

Segfault,

While I appreciate you taking time to respond, I have two points to raise with your response.

First, what version of OW are you talking about? What is your system configuration?

Second, at least under OW XE, the only way to change data selection permanently is to check out the version, change the data selection, and then check it back in. Even then, the data selection won't get up to the server until an update (or full) package is deployed to it. However, if you are talking about Processing Options, then yes, as soon as they are changed, the new values are used on both server and clients.

Hopefully this will clarify this matter.

 

[smorrison]

Don,
Actually Segfault is correct - the files under the SPEC directory in the Pathcode directory on the enterprise server are updated each time a client submits the version to it. If someone (as in this example) deletes the data selection whilst rinning the version and runs it on the server it will update these files.

If this happens you then need to submit version specs to the enterprise server from a client to restore these spec files.

 

[DSauve]

Stuart and Seg,

You are both absolutely right -- I was wrong, but I learned something. I ran a test and found that the data selection in the specs on the server does indeed get updated when a batch job is submitted on the server. Here's what I ran for a test, using a version of R014021 (Address Book Listing):

1. I submitted the Specs Only to the server to ensure a starting point.
2. I submitted the UBE from Batch Versions and got 365 names using the original data selection.
3. I ran the UBE from the Scheduler and got 365 names.
4. I submitted the UBE from Batch Version and got 2809 names using no data selection.
5. I ran the UBE from the Scheduler and got 2809 names. I even tried this step for a different User ID, but with the same results.

So I've seen that the data selection on the server does indeed get updated on the enterprise server when a client submits the version to it.

Thanks for your reply, and may we never stop learning!

 

[vetheridge]

before running the report via scheduler, did somebody run the same version from a fat client pc. if somebody runs a custom ube it is submitted with specs. scheduler will then pick up the submitted specs as they are the most recent ones.

i have been caught out with that one a lot