Often the user will specify the index node as searchapi_host in a selection file.
As Synda runs, a failure downloading a file may lead to a new search and a retry at whatever url the search discovered.
The new search should use the index node specified in the selection file. These changes keep track of searchapi_host by including it in the appropriate file and transfer objects.
That change overlaps with an improvement in the prioritization of data nodes in sdnexturl.py. So this improvement is also included in this branch and pull request.
Often the user will specify the index node as searchapi_host in a selection file. As Synda runs, a failure downloading a file may lead to a new search and a retry at whatever url the search discovered. The new search should use the index node specified in the selection file. These changes keep track of searchapi_host by including it in the appropriate file and transfer objects.
That change overlaps with an improvement in the prioritization of data nodes in sdnexturl.py. So this improvement is also included in this branch and pull request.