Architecture of Falcon
← Back to MySQL University main pagePresenter: Ann Harrison and Jim Starkey Time: Thursday, May 17, 2007, at 6am PST = 9am EST = 15 CET = 16 EET Estimated length of Session: 01:30 to 2:00 h Scribe: Paul DuBois Attendees: To register for this Session - Please enter your name here: * HakanKuecuekyilmaz * Alexander Nozdrin * AntonyCurtis * Oleksandr Byelkin * Georgi Kodinov
[edit] Session notes
[edit] Questions posted for the Session
These questions were posted during the session. The answers are discussed in the voice recording.
- joro_uni: can a tablespace be spread on mutiple disks ?
- timour|uni: can different tablespaces be allocated on different disks?
- timour|uni: JimStarkey, does Falcon perform execution of MySQL requests via the native operations or directly via "Logical Storage" requests?
- timour|uni: This design makes it mostly impossible to make use of existing internal optimizations, right?
- How does the seral log relate to the binary log?
- The two are physically separate. Statements are written to the binary log before the the commit record is written to the serial log.
- Does the Gopher have to read data when writing information?
- Yes, it can do. But we try to keep the information in the Serial Log entirely within memory, so in theory the information should be in memory. But, occasionally, the information may need to be read from disk.
