|
|
| Re: DBA related help need. [message #56622 is a reply to message #56600] |
Sat, 12 April 2003 02:23  |
Muhammad Waqas Shabir
Messages: 26
Registered: April 2003
|
Junior Member |
|
|
Dear Debraj!
You have asked theoretical questions which require book reading! But for for your convenience let me tell you these in quick!
Distributed transaction is which is affecting distributed databases.
Example:
You have a big company whose one branch in india and one in england. the head office is in canada!
Now you have three separate databases in these threee locations which are started up or shutdown independently. From head office you want to know that what is the accumulative sales from both of my branches. This will require the sales record in both the databases. Also these have to be in the open mode (started) before you can query them.
When you will run a query that will retrive record/info from these to separate databases then this query is called as distributed query/transaction. And the database said to be Distributed database i.e Physically separated nut logically one database.
----------------------------------------------------
Coming to your second question abat archive log mode.
See! we have two types of database modes when we talk about backup or recovery:
1: Archive Log mode (Completely Recoverable)
2: Non Archive Log mode (Incomplete recoverable (There is case where you can recover fully but that is rare and i wont discuss it right now))
Recovery Till the failure as evrery activity has been logged on the hard disk as well(Archiving):
In archive log mode we have backup of online redo log files on our hard disk which are called as archived log files. These files are used to store transactions which arestored in online redo log files. Now the question is why we need to store those online redo log files?
Well! the online redo log files are overwritten when the other redo log file gets full. This switching is called Log switching and the server process which does this is Log writer. The log writer dumps the online redo log file onto the arhiving file (ON Hardisk) and activates another server process called as Checkpoint. This process updates the headers of datafiles, control files, online redolog files with the current log number (LSN) Log sequence number.
If these files are not synchronized, the database wont open at all. So checkpoint is very very important process.
I think i havent confused you till now! Any ways good luck. Do read the ORACLE-DBA Track-Backup and recovery book. It will help you allllllllllot.
I am tired by now.I sould get some lunch! hahaha.
Good Luck
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
