> Here is the background: Anyone that is running a huge system like MARC > that has millions of uncompressed blob records in huge tables, needs to be > able to migrate, in real-time and without down-time, to compressed blobs. > Therefore, we need a way to know if a given field is compressed or not.
I hear you on that! We did the compression on the application end. When we started compressing all of the blobs in the table were uncompressed except newly added ones. We took advantage of the fact that zlib fails on decompression. So we wrote a function my_decompress() that takes the blob and decompresses it and if it fails just returns the original (assumed to be already decompressed). Works great and decompression gets divided among the webservers which scales better than having MySQL do it. However, you should develop a way to take tables offline. Lack of proper table maintenance can slow things down by a factor of 10 or more (and one of the reasons we can not use InnodDB). -steve-- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]