Database Status 10/19/01
- Postmortems:
- Daqcrl crashes: Dave Sarannen will install a new disk and copy files over.
- DBI: seems to work fine-congrats to Nick
- Plex: updates via complicated chain. Brian uses PHP interface to his db at Indiana, then loads new table descriptions into CVS, from which people pick it up and load it into their local dbs. DDM is working on improving this.
- Geometry: switch was thrown last week.
- DB maintenance and table distribution: DDM et al are working on this. After input from Nick, it will require a local database manager at each site to request new tables, which will come from the central data warehouse. The central warehouse will eventually be at FNAL, but temporarily elsewhere (probably www.physics.umn.edu? We considered daqcrl, but it crashes and there are security problems.) Updates will be incremental, though rewrites are not as bad as they sound. The plan is to allow db managers to pick groups of tables to update based on date and subject.
- DBU status: There have been requests to start putting DAQ pedestal data in the database- NW has come up with table designs. PMB has been looking through Sue's docs on making a dispatcher client. This will happen soon.
- DCS data status: very little has happened here. Needs work.
- BField and steel data: PMB has talked with Jeff about steel tables, and has some descriptions from him about data format, with some alpha versions of tables. BField?? -awaiting input-
- JobControl, Algorithm, Calibration -awaiting input-
- Factory Data:
- UMN factory is input through module 550 or so, (mapper data thru 503). Awaiting extrusion data for subsequent modules. Brian Sherwood has combined all the Java-based loaders into one program and added a nice GUI.
- ANL factory data is kept on paper. The php-web-based data entry system has been revived for them, TJM intends to enter the data.
- CIT factory data is kept on paper. Contact has been poor.
- Indiana has their own mysql-php based system. Their data is available, but seems quite happy where it is. We could bring it over to any mysql system very fast if there was any demand.
- UT-Austin: when last heard from (in August), they were re-designing their tables and needed to reanalyze their data. No decision on what to keep in the Minos experimental database.
- IU cable-test data is kept in an Access database. I have not seem much from them, but I dont believe there any troubles with it.
- C++ access to factory data: Leon has been charged with designing C++ access methods for the factory database. He is the only customer for this project, which makes it convenient. Casual users are encouraged to use MyODBC->MS Query->Excel, which works fine.
- RDBC, ODBC drivers, Oracle etc- No free ODBC drivers for linux turned up yesterday in a search through likely webpages, so we should continue to press CD for some Dennis Box time. Valeriy has been emailed, with no response yet. There has been no progress on Oracle recently, but it's still very much in the plan.
Pete Border, Oct 19, 2001