At this scale, your indexing job is prone to break in various ways.
If you want this to be reliable, it should be able to restart in the
middle of an upload, rather than starting over.
On 01/08/2013 10:19 PM, vijeshnair wrote:
Yes Shawn, the batchSize is -1 only and I also have the mergeScheduler
exactly same as you mentioned. When I had this problem in SOLR 3.4, I did
an extensive googling and gathered much of the tweaks and tuning from
different blogs and forums and configured the 4.0 instance. My next full run
is scheduled for this weekend, I will try with a higher mysql wait_timeout
value and update you the outcome.
--
View this message in context:
http://lucene.472066.n3.nabble.com/DIH-fails-after-processing-roughly-10million-records-tp4031508p4031779.html
Sent from the Solr - User mailing list archive at Nabble.com.
