We are running 5.0.2 348514c86@220530 on Windows servers, using RT index, SphinxQL to perform queries
We have encountered a problem where sometimes the server will stop responding to update requests, but will continue to respond to select. We then make a shutdown request to searchd, which fails in a ‘stopping’ state and we have to kill the process. We can start it again without any problem.
I am trying to see if I can replicate this issue. I don’t yet know if it is data or frequency of update requests causing it (I am preparing a load test)
So I’m just posting to see if anyone has seen similar behaviour and may have some insight?
Are there any of the fields returned in SHOW STATUS or SHOW THREADS that could give me any indication of the problem?
The search and query logs do not seem to have anything indicating failure