How many PATHs remain to be indexed?

[angelborroy]

Alfresco Search Services indexes PATH values from Repository using the CascadeTracker. Once the metadata and content has been indexed, some components (like the Enterprise Admin Web Console for the Enterprise version) will report that no indexation is happening. But CascadeTracker is still catching up with the Repo to get the paths indexed. This blog post will give you some recommendations to check the status of the PATH indexation.

When a Repository Node requires PATH re-indexing, a new property is added or modified in DB with the number of the Transaction that provoked this change (sys:cascadeTx).

The CascadeTracker, in SOLR side, gets a chunk of document transactions with status CASCADE_FLAG=1 (int@s_@cascade) and finds the nodes with that sys:cascadeTx equals to the transaction. After that, PATH properties are recalculated for the node and for every children requesting the information to the repository via the REST API. Once the work has been done, the transaction is marked back again to CASCADE_FLAG=0 in SOLR side to avoid processing it again with CascadeTracker.

The progress of CascadeTracker can be checked by finding all the pending transactions to be processed (CASCADE_FLAG=1) with a query similar to this one:

http://localhost:8983/solr/alfresco/select?fl=*,[cached]&indent=on&q={!term%20f=int@s_@cascade}1&sort=id%20desc&wt=json

• This is the number of transactions still to be processed by the CascadeTracker.
• When CascadeTracker work is done, the number of results for this query is equals to 0.

Are you using a different approach to monitor CascadeTracker progress? Share it in the comments below!