The tab health checker shows an overview of all the build-in checks which are available for the Archive.
Running a health check is always a two step process.
- Run the analysis. In this step the health check will verify if a problem exists in your archive. To run the analysis for a certain health check simply press the analyze button for a specific health check. After the analysis is done the status field will show how many (if any) problems were found. More detailed information can be found by downloading the report. If the analysis has found problems the download button in the report column will be enabled and will allow the user to download a file containing more detailed information about the problems found.
- Repair of detected problems. To start the repair action (this can only be done after the analysis has run and found problems!)
- press the Repair button in the repair column.
- Reviewed the results / changed files in the archive.
- If needed, fallback to the situation prior to the repair. In the root folder of the archive a folder HealthChecker will be created. This folder will contain a backup of the files which were changed during the repair action. The files are copied with their relative path to the data folder. A meta data file in folder /opt/fews/archive/data/2017/01/area/01/ will be backuped to folder /opt/fews/archive/HealthChecker/<health check id>-<datetime)/2017/01/01/area/01.
- Run the harvester to update the catalogue with the changes made in your meta data files.
Available health checks
Missing source ids in observed data sets
It is possible to assign a source to a netCDF-file containing observed data. In some cases some netCDF-files don't have a source id assigned while other netCDF-files with the same file name have a source id assigned. This can happen if, for example, the source id was added to the data export while the data export has already been running operationally for some time. In this case the already exported netCDF-files won't have the source id assigned while the newly exported files will have source id assigned. This tool will determine first how the netCDF-files are mapped to source ids. If the netCDF-files with file name A always have source id X assigned or don't have a source id the following will be done. The netCDF-files without a source id will automaticly have source id X assigned. This will be done by added the source id to the meta data files.
Search for orphan netCDF files (external forecasts)
At the end of 2017 the following bug was found in the data export for external forecasts. External forecast will have an incorrect meta data file if the data set contains more than one netCDF-file. The bug in export caused that only the last exported netCDF-file was listed in the meta data file. This health check will detect this problem and fix it. This health check only works for scalar time series at this moment.
