Re: PROPOSAL mysqldump change for InnoDB

2001-05-17 Thread Heikki Tuuri
Thu, 17 May 2001 15:53:59 >To: [EMAIL PROTECTED] >From: Heikki Tuuri <[EMAIL PROTECTED]> >Subject: Re: PROPOSAL mysqldump change for InnoDB > >Steve and Monty, > >yes, Monty is right, with > >mysqldump --quick > >you should get much faster import speed. Bu

RE: PROPOSAL mysqldump change for InnoDB

2001-05-17 Thread Andre . Konopka
Hi, Yes, sounds great, it is possible to implement it in the next version Andre On 17-May-01 Michael Widenius wrote: > > Hi! > >> "Andre" == Andre Konopka <[EMAIL PROTECTED]> writes: > > Andre> Hi Michael, > Andre> maybe something 'off topic'... > > Andre> I think a option to set

RE: PROPOSAL mysqldump change for InnoDB

2001-05-17 Thread Michael Widenius
Hi! > "Andre" == Andre Konopka <[EMAIL PROTECTED]> writes: Andre> Hi Michael, Andre> maybe something 'off topic'... Andre> I think a option to set 'autocommit=0' is needed. Andre> Last week I have made some data manipulation, stupidly my 'where condition' was Andre> not correct. After pre

Re: PROPOSAL mysqldump change for InnoDB

2001-05-17 Thread Heikki Tuuri
Steve and Monty, yes, Monty is right, with mysqldump --quick you should get much faster import speed. But I have to correct Monty: a big rollback segment does not slow down import considerably. Regards, Heikki At 12:25 AM 5/17/01 +0300, you wrote: > >Hi! > >> "Steve" == Steve Ruby <[EMA

RE: PROPOSAL mysqldump change for InnoDB

2001-05-16 Thread Andre . Konopka
Hi Michael, maybe something 'off topic'... I think a option to set 'autocommit=0' is needed. Last week I have made some data manipulation, stupidly my 'where condition' was not correct. After pressing 'return', mysql returns with '4576 rows affected'. Not that was I expected... I didn't set au

PROPOSAL mysqldump change for InnoDB

2001-05-16 Thread Michael Widenius
Hi! > "Steve" == Steve Ruby <[EMAIL PROTECTED]> writes: Steve> Due to the fact that piping a table from mysqldump to an INNODB table Steve> can be as much as 30X slower if the insert is not done with Steve> transactions, Steve> would it make sense to add some option to mysqldump to make eve

Re: PROPOSAL mysqldump change for InnoDB

2001-05-15 Thread Sinisa Milivojevic
Steve Ruby writes: > > > Due to the fact that piping a table from mysqldump to an INNODB table > can be as much as 30X slower if the insert is not done with > transactions, > would it make sense to add some option to mysqldump to make every X > inserts a transaction. > > for example > > mysqld

PROPOSAL mysqldump change for InnoDB

2001-05-14 Thread Steve Ruby
Due to the fact that piping a table from mysqldump to an INNODB table can be as much as 30X slower if the insert is not done with transactions, would it make sense to add some option to mysqldump to make every X inserts a transaction. for example mysqldump --transactionrows=10 mydatabase |