>>>>> "Tim" == Tim Bunce <Tim.Bunce@stripped> writes:
Tim> On Mon, Jan 10, 2000 at 01:18:00PM -0500, Bill Marrs wrote:
>> >The command is, as I said: cd your_database; myisamchk -fe *.MYI
>> -f will restart with -r if it finds there are any errors in the table. My
>> desire for this recovery to work the first time without a need for someone
>> to re-run it (the --help for -o says "Slower than '-r' but can handle a
>> couple of cases where '-r' reports that it can't fix the data file"), made
>> me think that using -ef might not always fix everything.
>> Also, why isn't the functionality of -A (auto-increment repair)
>> incorporated into -r and -o?
Tim> And why doesn't -r auto restart myisamchk with -o? (Or why isn't there
Tim> an option to do that?)
Simply because in normal use -o should never be needed, except if
there is some bugs in the sort code that builds the indexes; Nowadays
the sort code is pretty good tested and should be as good (or even
better) than -o.
Tim> Seems to me that isamchk/myisamchk has not been designed for unattended
Tim> use. But that's what many people need, and what people used to the
Tim> crash-proofness of transactional databases would expect.