Hello SuSE folkz, I'm sorry for these strange questions After reading "Web Performance Tuning" 2nd Edit. 2002 by Patrick Killelea Pub.O'Reily. ISBN 0-598-00172-x I have the following assumptions: Linux kernel uses buffered IO file access when it's reading and writing to the physical partition/device with its native FS. Linux kernel uses unbuffered IO file access if a program (e.g. DB engine) reads and writes to the raw partition/device. I assume it's because DB engine handles its data transaction on raw device/partition without going through the kernel IO calls. Does all above apply to the case if drives support DMA? Why it's nessesary to bind raw devices to block devices ( bind /dev/raw/raw1 to /dev/hdb1 ) if a database engine reads and writes to the raw partition without such binding? Will this binding improve performance of the DB engine or this needs to be done only in order to read from Linux what has been written on a raw device? Thank you in advance for any answers or sources of further information. Alex
* Alex Daniloff (alex@daniloff.com) [021009 14:27]:
Does all above apply to the case if drives support DMA?
DMA? If you are using ide drives and you are concerned about performance the first step should be to replace them with fast scsi disks.
Why it's nessesary to bind raw devices to block devices ( bind /dev/raw/raw1 to /dev/hdb1 ) if a database engine reads and writes to the raw partition without such binding?
raw1 is a character device that needs to be bound to a real disk. -- -ckm
participants (2)
-
Alex Daniloff
-
Christopher Mahmood