Thanks so much Serge for for accurate remarks, I sure this will help Fabio getting some nice hardware :-) On 02/18/2016 01:39 PM, Serge Monkewitz wrote: > > I said: >> That link says that the page merge threshold used to be hardcoded at >> 50% free, hence my worst case estimate. It’s plausible that this is >> way too pessimistic in practice, and that sorting before loading >> results in relatively full pages. I just don’t know for certain. > > I looked around some more, and based on > https://www.percona.com/files/presentations/percona-live/london-2011/PLUK2011-b-tree-indexes-and-innodb.pd (the > in-order INSERTs slides), it seems that loading data that is > pre-sorted by PK results in very good page fill percentage. So 2TB > should be more than enough. > > > ------------------------------------------------------------------------ > > Use REPLY-ALL to reply to list > > To unsubscribe from the QSERV-L list, click the following link: > https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=QSERV-L&A=1 > ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the QSERV-L list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=QSERV-L&A=1