Closed (fixed)
Project:
iPaper
Version:
6.x-1.x-dev
Component:
Miscellaneous
Priority:
Normal
Category:
Bug report
Assigned:
Unassigned
Reporter:
Created:
29 Jun 2009 at 20:36 UTC
Updated:
5 Jul 2009 at 17:06 UTC
I have a series of PDF documents I'm uploading to Scribd through iPaper. Any document I upload and publish causes my search index to fail.
Any ideas? Each document is about 40 pages, so they aren't super huge or anything.
Comments
Comment #1
rares commentedDoes it work with very small documents (and no large documents)?
The only way ipaper cron hook might be causing this is if you added a large number of large documents and the Scribd server is slow. Can you enable logging in the module administration and check if there are API calls with method getDownloadURL the first time you run cron after the upload. If cron fails and they appear every time you run cron, then it must be that they are not completing and the module is not able to get the full text from scribd in due time.
Comment #2
djudd commentedI'm getting no logs at all on this. Cron runs, and even says it completes, but I always have a few items that did not index.
For example, I've unpublished all of my iPaper items except one, and I get the following message in the admin/settings/search page:
98% of the site has been indexed. There is 1 item left to index.
Anything nodes I create after this won't index because they get stopped at the iPaper node. So to index my normal story nodes, I have to un-publish all iPaper documents, re-index my site using another module, and then re-publish the iPaper nodes.
Now, I have noticed that iPaper times out every time I upload a new document to Scribd through the module, so I'm taking a wild guess that you may be right on that issue. My hosting company has site the max timeout to 45 seconds, and I can't change that.
In this very rare case, is there a way to tell iPaper nodes not to be indexed by the search module? I don't really need them to be searchable anyway, and since I have to upload a new document daily, this is going to be disastrous down the road for me.
I've tried using Search Config to disable indexing of the node, but it doesn't work, they still crap out my search index every day.
Comment #3
djudd commentedI've found the answer. I needed to disable full text inclusion in the iPaper module.
Since I'm building a newspaper website, I guess all of the stories being uploaded in our digital edition was a little much to be indexed. Search is working fine now as best as I can tell.