Issue Number | 4382 |
---|---|
Summary | Re-publish Full Document Type |
Created | 2018-01-11 16:49:11 |
Issue Type | Inquiry |
Submitted By | Englisch, Volker (NIH/NCI) [C] |
Assigned To | Kline, Bob (NIH/NCI) [C] |
Status | Closed |
Resolved | 2018-10-16 08:34:50 |
Resolution | Cannot Reproduce |
Path | /home/bkline/backups/jira/ocecdr/issue.219689 |
When running a Re-publish job for a complete document type with a
limited number of documents, i.e. DrugInfoSummaries, the publishing
works without issues.
However, if the number of documents of the selected document type is
large - in the order of several thousand like Media - the publishing job
fails.
At first, all of the documents to be published are selected and the
publishing job is submitted. The publishing continues even though an
error message gets created. The error message created is
"Failure: republish(): Job 15276 of this publication type is still pending"
The Job 15276, however, is the job created by this re-publish attempt itself.
What should be happening is a message indicating "Job15276 successfully submitted" instead of submitting an error message.
This is the content of the Re-publish email notification submitted as part of the "failure":
Failure republishing CDR documents:
republish(): <Errors><Err>Job 15276 of this publication type is still pending</Err></Errors>
Can you help me reproduce this behavior tomorrow, ~volker?
The system is likely working too fast when you're in Arlington. That's why you can't reproduce it. :-)
I went back through the publishing logs on PROD going back to the beginning of this year, and the only incident recorded for the condition of a publishing job of the requested type already being in the queue was legitimate (a mailer requested by Bonnie right after submitting a mailer request which was still in the queue). If there is more recent evidence which I haven't found, please re-open the ticket and attach that evidence.
We almost never will run the Re-publish jobs on PROD. The republish
interface has been created so we can republish documents to the lower
tiers where pushing would be suppressed because we're only pushing
changed documents by default.
I haven't had the need to publish a full doctype in quite some time so
I'm not surprised you're not finding evidence in the log files at this
time.
As a test I started our largest two document types (Terms and Media) on DEV without any issue. I'm OK to keep this one closed.
Elapsed: 0:00:00.000684