On Thu, 15 Nov 2001 23:29:27 -0800, Jeremy Zawodny used a few recycled electrons to form: | On Thu, Nov 15, 2001 at 04:14:58PM -0600, Michael Brunson wrote: | > | > Has anyone ran any benchmarks as to what affect different number of | > RAID_CHUNKS has on speed? | | The more chunks, the slower your retrieval speed is likely to be. But | how many chunks do you really need? More than a few?
The .MYD right now is right at 2G, so 2 chunks would work. We would need to rebuild again once the combined size hits 4G. I could see the table growing, but probably not more than another 10 million rows in the next year. (It's at 30M now.) | > mysql> select count(*) from names where sld like | > 'foo%'; | > +----------+ | > | count(*) | | > +----------+ | > | 56033 | | > +----------+ | > 1 row in set (0.33 sec) | | Is there an index on "sld"? It would seem so. If that's the case, | the test is relatively meaningless. The index will be used to answer | the query, not the raid'ed data files. Yes, sld is the primary key for the table. In fact, sld is the only column in the table. | > mysql> select count(*) from names where sld like | > '%foo%'; | > +----------+ | > | count(*) | | > +----------+ | > | 151460 | | > +----------+ | > 1 row in set (2 min 3.70 sec) | | Same here. | | If you have a lot of "%foo%" queries, you probably want to look at | fulltext indexing. Any ideas why the times would jump up for 3 chunks and back down for 5 and 15? Anyone think this might be more of a memory/swap issue? Thanks for any input, Michael --------------------------------------------------------------------- Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
