mysql IDC之間網(wǎng)絡不穩(wěn)定導致出錯分析，大家有遇到過沒

helbreathszw

因為數(shù)據(jù)庫主庫的IDC和幾個relay從庫的IDC之間網(wǎng)絡不穩(wěn)定或是網(wǎng)絡嚴重丟包情況時有發(fā)生。然后就會出現(xiàn)sql語句拼裝出錯，其他在和主庫同IDC的從庫就沒問題，很莫名，我的數(shù)據(jù)庫版本是Server version: 5.0.41 MySQL Community Server (GPL)，不知道有其他人遇到過沒？


071218  5:52:09 [ERROR] Slave: Error 'You have an error in your SQL syntax; check the manua
l that corresponds to your MySQL server version for the right syntax to use near 'UPDATeek_
count=read_week_count+2, read_month_count=read_month_count+2 WHERE book' at line 1' on quer
y. Default database: 'vipbook'. Query: 'UPDATeek_count=read_week_count+2, read_month_count=
read_month_count+2 WHERE book_id=43113�', Error_code: 1064

helbreathszw

環(huán)境是centos4.4,發(fā)生sql拼裝是大概網(wǎng)絡不穩(wěn)定時候，會有1，2次出現(xiàn)，不是頻繁。

[ 本帖最后由 helbreathszw 于 2007-12-21 13:16 編輯 ]

helbreathszw

找到問題所在了http://bugs.mysql.com/bug.php?id=26489
For the general issue of corruption of data in transit see bug #25737. We know that TCP's
CRC checks aren't good enough to protect the integrity of the data being transferred over
replication links and that corruption can happen and produce changed characters even with
a valid CRC.

This specific problem may not be due to that, unless the ending of the logs coincides with
a broken connection, which may have a period of bad connectivity and errors preceding it.