[Xapian-discuss] Searching log files?

Kevin Duraj kevin.softdev at gmail.com
Tue Jan 19 20:35:52 GMT 2010


I have only one hard disk and little memory that is why I use Xapian
on my own server.  Xapian runs very smoothly, stable  and fast that is
why I recommend to use it only on your own server.

Lucene on other side is much better when you building search engine
for someone else, because is slow, leaking memory, crashing left and
right, Lucene 3.0 is not compatible with Lucene 2.4 and many other
related problems will keeps you employed. You will be always working
your time with Lucene index optimization, merging, debugging trying to
figure out why indexing and searching is so slow. It is a great search
engine to keep you employed.

On the other hand I developed several Xapian large search engines,
they never break, I lost my job, all the Xapian search engines still
running smoothly long time after I am gone, so they never call me to
fix it, without anybody know how they run and that is not very good if
you want to be employed.

Kevin Duraj
http://find1friend.com/


On Tue, Jan 19, 2010 at 3:59 AM, Paul Boddie <paul.boddie at biotek.uio.no> wrote:
> Kevin Duraj wrote:
>>
>> Moazam,
>>
>> ... and here is an example how Xapian perform searches on 100 millions
>> records using one physical 1 terabyte hard drive ...
>>
>
> How big is the index? How much memory do you have? Is the searching done in
> parallel or on one big index? I noticed that if you enter "all my" as the
> search term, you get an internal server error - this isn't so surprising
> since these terms, if you index them, would occur a lot. It looks like
> you're doing a phrase query with the search terms - is that correct?
>
>> If you do not need so much power, you might also consider to use less
>> powerful
>> search engine like Lucene.
>>
>
> I think it's inaccurate to portray Lucene as less powerful, especially given
> that it's used on some large/Internet-scale search problems. Certainly,
> Lucene has a number of advantages over Xapian - space efficiency in position
> storage being one I noticed the most - and the converse is also true.
>
> Paul
>



More information about the Xapian-discuss mailing list