Expert use only - the default configuration should “just work”. The options described here give precise control over the pipeline workflow for special cases.
When the rover retrieve command is used, the following steps are taken:
Indexing - Any modified files in the repository are indexed.
This step can be omitted with --no-pre-index.
Availability - The Availability service is contacted to see what
data are available. This is done by the retrieve command itself.
Comparison - The available data re compared with the local index
to see what timespans are missing. Again, this work is done by
the retrieve command.
Chunking - Required timespans are split and arranged by N_S_L_C
and day. This work is done by the Download Manager as part of the
retrieve command.
Separate Processes - The following steps are done in a separate process for each download. This reduces the damage that any error can have on the retieval as a whole.
Download - A day’s data for a particular N_S_L_C are
downloaded from the Data Select service. This is done by the
rover download command.
Indexing for Ingest - The mseedindex command is used to
index the downloaded data. This is done by rover ingest. This
and later steps can be omitted with --no-ingest.
Ingest - Each block of indexed data is added to the
repository. This is done by rover ingest, which takes the byte
offset and length of the block and appends that data to the
existing data in the repository.
Indexing - The command rover index updates the index for the
repository. This step can be omitted with --no-index.
Summary - A summary table is generated. This step can be ommitted
with --no-post-summary.
The options --no-ingest and --no-index can be used to stop the
separate processes short.
Initial indexing can be omitted with --no-pre-index and final
summarizing with --no-post-summary.
These options also affect low-level commands thare are used
individually. So using --no-index with rover ingest would mean
that rover index was not called after ingesting data.
