I have recently started to fill my wiki with lots of content. Currently
I have 4772 XWiki documents and 2260 XWiki objects. Not so much I think,
but it looks like I'm reaching a limit.
This is first time I have looked into SQL internals of XWiki.
Goddamn! There were no indices except primary ones! OK, I have lowered
load time from 15sec to 9sec having created indices (e. g. XWD_FULLNAME
in xwikidocs)
Still too much.
Maybe I could improve it any further, but some queries are maid the way
I don't know obvious way to make index for them. For instance, I can see
"where hidden <> 1 or hidden is null" in queries. It would be nice to
see "where hidden = 0" instead (and null being not allowed)
Xen:~ OCTAGRAM$ time GET
http://blog.toom.su/
[...snip...]
</div></div></body>
</html>
real 0m8.395s
user 0m0.158s
sys 0m0.052s
Xen:~ OCTAGRAM$ time GET
http://pascal.toom.su/Index
[...snip...]
</div></div></body>
</html>
real 0m5.537s
user 0m0.157s
sys 0m0.043s
Xen:~ OCTAGRAM$ time GET
http://www.xwiki.org/xwiki/bin/view/Main/News
[...snip...]
</div></div></body>
</html>
real 0m1.981s
user 0m0.152s
sys 0m0.034s
2 second is also too much, but how do you keep response time that low?
Is it just due to the fact you scatter documents across different
databases or may be there was a huge improvement from 2.4M1?
http://pascal.toom.su/Index is even not a dynamic page, it just has lots
of wiki links.
--
If you want to get to the top, you have to start at the bottom