Hi Robert,
Are you missing exactly 1 out of every 50k documents? Or is it just
something close to that number?
Are you able to identify the specific records that aren't making it
into the index?
These kinds of things are typically due to something like two row keys
being mapped to the same Solr document id, although this isn't always
easy to find. If you don't have specific information about the actual
records causing this problem, I would try one (or both) of the
following:
1. enable DEBUG logging for hbase-indexer and see if there are record
updates that are being unexpectedly not sent to Solr (or something
related)
2. follow the metrics that are published via JMX while you are
ingesting data into HBase -- these metrics provide information on
counts of record that make it through various stages in the indexing
pipeline, so should help determine where a record is stopping if it
isn't being indexed
- Gabriel
> --
> You received this message because you are subscribed to the Google Groups
> "HBase Indexer Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to
hbase-indexer-u...@googlegroups.com.
> For more options, visit
https://groups.google.com/d/optout.