Re: reasonable checkpoint interval

Kahler, Russell P. wrote: > > }I disagree with the idea of increasing LRU writes and decreasing Chunk > }writes. > }The porblem with this is that Chunk writes are much faster then LRU writes. > } Chunk > }writes will be much faster since they will not be competing with other > }threads to do their work. Also Chuck writes work in sequence with the disk > }to shorten the amount of head movement while performing these writes. > } > } > }I think you need to balance LRU writes to allow the most amount of Chunk > }writes while not increasing forground writes and not increasing Check point > }time to an unexceptable time. I agree, chunk writes are sorted into disk order and written using big buffers. These are the most efficient writes, but a checkpoint will effectivily halt user activity, therefore a prolonged checkpoint is noticed by users. Whereas LRU writes, although less efficient are effectivily run in background, and are not so obvious to users. You should be looking for a balance between the two, with no foreground writes. -- Mark. +-------------------------------------------------------------------------+ |Mark D. Stock - The West Solutions Group http://www.west.co.za| | | |Email: marks@west.co.za +------------------------------------------------+ |Tel: +27 11 803 2151 |If it doesn't work... force it! | |Fax: +27 11 803 2189 |If it breaks... it needed replacing anyway! | |Cell: +27 83 250 2325 |Well, that's how I code anyway! | +------------------------+------------------------------------------------+