Frequently Asked Questions

Q: I am completely new to databases, what should I do first?
A: Please, read this FAQ list, and database API documentation :
Database documentation
Then, please read You do not have access to view this node
Don't forget to log in, most of the information is STAR-specific and is protected; If our documentation pages are missing some information (that's possible), please as questions at db-devel-maillist.

Q: I think, I've encountered database-related bug, how can I report it?
A: Please report it using STAR RT system (create ticket), or send your observations to db-devel maillist. Don't hesitate to send ANY db-related questions to db-devel maillist, please!

Q: I am subsystem manager, and I have questions about possible database structure for my subsystem. Whom should I talk to discuss this?
A: Dmitry Arkhipkin is current STAR database administrator. You can contact him via email, phone, or just stop by his office at BNL:
Phone: (631)-344-4922
Email: arkhipkin@bnl.gov
Office: 1-182

Q: why do I need API at all, if I can access database directly?
A: There are a few moments to consider :
    a) we need consistent data set conversion from storage format to C++ and Fortran;
    b) our data formats change with time, we add new structures, modify old structures;
    b) direct queries are less efficient than API calls: no caching,no load balancing;
    c) direct queries mean more copy-paste code, which generally means more human errors;
We need API to enable: schema evolution, data conversion, caching, load balancing.

Q: Why do we need all those databases?
A: STAR has lots of data, and it's volume is growing rapidly. To operate efficiently, we must use proven solution, suitable for large data warehousing projects – that's why we have such setup, there's simply no subpart we can ignore safely (without overall performance penalty).

Q: It is so complex and hard to use, I'd stay with plain text files...
A: We have clean, well-defined API for both Offline and FileCatalog databases, so you don't have to worry about internal db activity. Most db usage examples are only a few lines long, so really, it is easy to use. Documentation directory (Drupal) is being improved constantly.

Q: I need to insert some data to database, how can I get write access enabled?
A: Please send an email with your rcas login and desired database domain (e.g. "Calibrations/emc/[tablename]") to arkhipkin@bnl.gov (or current database administrator). Write access is not for everyone, though - make sure that you are either subsystem coordinator, or have proper permission for such data upload.

Q: How can I read some data from database? I need simple code example!
A: Please read this page : You do not have access to view this node

Q: How can I write something to database? I need simple code example!
A: Please read this page : You do not have access to view this node

Q: I'm trying to set '001122' timestamp, but I cannot get records from db, what's wrong?
A: In C++, numbers starting with '0' are octals, so 001122 is really translated to 594! So, if you need to use '001122' timestamp (any timestamp with leading zeros), it should be written as simply '1122', omitting all leading zeros.

Q: What time zone is used for a database timestamps? I see EDT and GMT being used in RunLog...
A: All STAR databases are using GMT timestamps, or UNIX time (seconds since epoch, no timezone). If you need to specify a date/time for db request, please use GMT timestamp.

Q: It is said that we need to document our subsystem's tables. I don't have privilege to create new pages (or, our group has another person responsible for Drupal pages), what should I do?
A: Please create blog page with documentation - every STAR user has this ability by default. It is possible to add blog page to subsystem documentation pages later (webmaster can do that).

Q: Which file(s) is used by Load Balancer to locate databases, and what is the order of precedence for those files (if many available)?
A: Files being searched by LB are :

1. $DB_SERVER_LOCAL_CONFIG env var, should point to new LB version schema xml file (set by default);
2. $DB_SERVER_GLOBAL_CONFIG env. var, should point to new LB version schema xml file (not set by default);
3. $STAR/StDb/servers/dbLoadBalancerGlobalConfig.xml : fallback for LB, new schema expected;

if no usable LB configurations found yet, following files are being used :

1. $STDB_SERVERS/dbServers.xml - old schema expected;
2. $HOME/dbServers.xml - old schema expected;
3. $STAR/StDb/servers/dbServers.xml - old schema expected;