thomasmueller commented on code in PR #2469:
URL: https://github.com/apache/jackrabbit-oak/pull/2469#discussion_r2297461776


##########
oak-core/src/main/java/org/apache/jackrabbit/oak/plugins/index/importer/ClusterNodeStoreLock.java:
##########
@@ -38,10 +38,14 @@
  */
 public class ClusterNodeStoreLock implements 
AsyncIndexerLock<ClusteredLockToken> {
     /**
-     * Use a looong lease time to ensure that async indexer does not start
-     * in between the import process which can take some time
+     * Use a long lease time to ensure that async indexer does not start
+     * in between the import process which can take some time.
+     * For a very large repository (about 2 billion nodes) the longest
+     * index import can take 2.5 hours. 6 hours has then a safety margin of 
100%
+     * above that. A diff of 6 hours is typically OK with the document node 
store,
+     * but beyond that it gets harder.
      */
-    private static final long LOCK_TIMEOUT = TimeUnit.DAYS.toMillis(100);
+    private static final long LOCK_TIMEOUT = TimeUnit.HOURS.toMillis(6);

Review Comment:
   There's also cost of making things configurable. And how could it be 
configured, and where would the configuration be?
   
   6 hours timeout would mean there is an indexing job where the import alone 
takes more than 6 hours... I actually can't imagine such a case, because that 
would mean indexing takes multiple days. (The risk of the indexing process to 
die on it's own is very large in this case... but let's ignore that.) Let's 
assume there is such a case, eg. manually run, on a separate machine... In this 
case, the async indexing can be stopped.
   
   So, for this extreme case, there are already two workarounds: 
   * stop the async indexing.
   * create a custom Oak version with increased timeout.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: oak-dev-unsubscr...@jackrabbit.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org

Reply via email to