Wednesday, July 8, 2015

ADOP failed jobs

From AD.4 and TXK.4 , autoskip option for failed jobs is turned off. -- (flags=noautoskip).

Before this, autoskip was enabled (flags=autoskip) and so autoskip.log needs to checked after patch application and before cutover.


From Ad.4, a new option wait_on_failed_job is available to deal with failed jobs.

This will allow the user to use adctrl to update the workers and jobs when a job has failed.

The option is valid only for SQL failures anyhow.


In the case of failures during the "generate" portion of a patch,
i.e Forms, Reports etc the generation code(fs actions like genpll etc) does not obey the wait_on_failed_job parameter.
As with previous EBS releases, if form or report generation failed the worker returns success
and then populates the list of failures in a text file.

This is then read by manager code and the manager throws the prompt:

***
Continue as if it were successful :
***


There is no token entry in a defaults file for prompt “Continue as if successful”.
 The autoskip flag was used to pipe a "no/yes" value for ALL these prompts.
 Currently there is no flag or option to address individual failed prompts.
 The flags=noautoskip is now the default with AD.Delta 4 and TXK Delta 4, therefore a response of "no" is passed to these prompts
 so the patch session terminates as follows:

AutoPatch could not find a response to the above prompt
or found an incorrect response in the defaults file.


You must run AutoPatch in an interactive session
and provide a correct value.

No comments:

Post a Comment