[TYPO3-english] Search Engine problems
Jens W. Skov
JS at wexoe.dk
Wed Feb 10 14:55:33 CET 2010
I'm new on this list. I manage a Typo3 site that I have taken over from someone else.
I have some difficulties making the search engine work properly.
Sometimes I have MySQL taking up 100% cpu in periods. It seems to be caused by a query ny tx_indexedsearch.
# Query_time: 4599 Lock_time: 0 Rows_sent: 2549 Rows_examined: 87553393
SELECT ISEC.*, IP.*, MAX(IR.flags) AS order_val1, SUM(IR.freq) AS order_val2
FROM index_words IW,
(IW.metaphone = 0) AND
IP.phash IN (108783,131499,... en masse tal.... ,268364100)
AND ISEC.phash = IR.phash
AND IP.phash = IR.phash
AND ISEC.page_id IN (404,432...... tal.... ,464,1)
GROUP BY IP.phash,ISEC.phash,ISEC.phash_t3,ISEC.rl0,ISEC.rl1,ISEC.rl2 ,ISEC.page_id,ISEC.uniqid,IP.phash_grouping,IP.data_filename ,IP.data_page_id ,IP.data_page_reg1,IP.data_page_type,IP.data_page_mp,IP.gr_list,IP.item_type,IP.item_title,IP.item_description,IP.item_mtime,IP.tstamp,IP.item_size,IP.contentHash,IP.crdate,IP.parsetime,IP.sys_language_uid,IP.item_crdate,IP.cHashParams,IP.externalUrl,IP.recordUid,IP.freeIndexUid,IP.freeIndexSetId
ORDER BY order_val1 DESC,order_val2 DESC;
Can anyone point out what it might be caused by?
I have set up crawler now and completely rebuild the index, but both now and before I get strange results. Fx. If I search for 'an200' I get a lot of pdf results that does not contain 'an200'. I can get the same result with other bogus search strings, but I can also search for valid strings that does exist and they will show up fine.
Could it be related to pdftools? I notice that it gives a warnings that it does not support pdf 1.6.
Right now i'm upgrading to Typo3 4.3 and after that I'm considering switching to mnoGoSearch or Google.
Do you have any recommendation or warnings?
Jens W. Skov
More information about the TYPO3-english