IBM Connections search problems with ccm
one of our daily jobs (ok, that`s a bit overdone 😉 ) is to handle IBM Connections search problems…
No to be honest, it happens quite often that customers upload files and data that contains “bad data” – or “bad, non readable entries”. Normally I would expect that the apache lucene search engine simply skips these entries and does not add them to the search index. But… it sometimes does not, what results in an incomplete or non-refreshing searchindex.
It is really tough to find the erroneous document then – you have to go through all the seed lists and find the entry that causes the problem.
This time I had a very special case… The customer did upload files into CCM and used its own document types for this. But when searching for the files, the result looked like this:
Normally you should see the matching document type instead of its ID.
Well I was sure; this has to do with the search index that has a problem (which was comfirmed by looking into the log):
Error: CLFRW0932E: the document identifier could not be retrieved from the ECM-Server ...
While retrieving the document types list, an internal error was returned.
When things like this happen, I first of all try to open the URL manually in my browser to see the result:
When entering this URL, I go a Filenet error, that there is a “configuration problem exists”!
I tried to narrow down, which document types work and which do not by entering:
https://SERVERNAME/dm/atom/library/allobjectstores/documenttypes/feed?pageSize=1&page=1 https://SERVERNAME/dm/atom/library/allobjectstores/documenttypes/feed?pageSize=2&page=1 https://SERVERNAME/dm/atom/library/allobjectstores/documenttypes/feed?pageSize=3&page=1 ...
I got results for these and I noted down the document type names
Then I looked into the acce admin interface for the document types that are configured:
These are the available documents:
I compared these with the list I generated while calling the atom feed for the document types list…
AND VOILÁ… I found a custom document type “Document BLABLA” (the one in the red frame). I asked the customer to delete it, as it was not used anywhere…
Now I`m able to call the URL that failed before:
Another manual rebuild of the ecm search index and the error was solved (Sorry, I have to “grey-out” a lot):
It`s rather hard to see but the document type is now displayed again 😉 yeah.
To be honest, I do not know, why this document type contained errors! But as it was not used by anyone, it was a faster and smarter way to delete it rather than opening a PMR for this.