>>>>> "Yohanes" == Yohanes Santoso <ysantoso@jenny enusa net>
> <ysantoso@stripped> writes:
Yohanes> Right now, we have almost 60 millions row in a table. So far MySQL is doing
> great. However, in a couple of months, once our research project is approved, we are going
> to store way more than 60 millions row. Looking at the stat from myisamchk below, I don't
> think there will be any problem on data file length.
Yohanes> But I am concerned about the max index file size. Right now, the only thing
> indexed is the auto-incremented primary key. We are planning to have more columns indexed.
> What if we needs more than 17G of index? Will there be a way to change the max index size
> later, or should we be ready to separate the data into more than one table?
Yohanes> ai20:~/var/rts>myisamchk -dvi queue.MYI
Yohanes> MyISAM file: queue.MYI
Yohanes> Record format: Packed
Yohanes> File-version: 1
Yohanes> Creation time: 1999-10-05 9:24:14
Yohanes> Status: open
Yohanes> Auto increment key: 1 Last value: 56293811
Yohanes> Data records: 56293811 Deleted blocks: 0
Yohanes> Datafile parts: 56293811 Deleted data: 0
Yohanes> Datafile pointer (bytes): 6 Keyfile pointer (bytes): 3
Yohanes> Datafile length: 4920206844 Keyfile length: 851499008
Yohanes> Max datafile length: 281474976710654 Max keyfile length: 17179868159
Yohanes> Recordlength: 263
Yohanes> table description:
Yohanes> Key Start Len Index Type Rec/key Root
Yohanes> 1 1 4 unique unsigned long 1 234540032
When creating a table, MYSQL calculates how big data file pointer and
key file pointer it will need to satisify the create parameters.
Both the Datafile pointer and the keyfile pointer may be up to 7
bytes long (which should be enough for at least a while).