Lance,
I have seen this error when the Solr process hit the maximum file
descriptors (because the commit triggered an optimize). Make sure your
maxfds is set as high as possible. In my case, 1024 was not nearly
sufficient.
--Casey
On 10/19/12 6:20 PM, Lance Norskog wrote:
When a transient error happens during an autocommit, the error does not cause a
safe rollback or notify the user there was a problem. Instead, there is a write
lock failure and Solr has to be restarted. It run fine after restart.
Is this a known problem? Is it fixable? Is it unit-test-able?
