Re: "Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-07 Thread mos
th 100g of disk space dedicated to the sort file. It started using "Repair with Filesort" for the first 7 hours, then it switched to "Repair with KeyCache". It should only use one of the Repair methods. If it switched then something was wrong with Repair by filesort (maybe ou

Re: "Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-06 Thread Harrison Fisk
t file. It started using "Repair with Filesort" for the first 7 hours, then it switched to "Repair with KeyCache". It should only use one of the Repair methods. If it switched then something was wrong with Repair by filesort (maybe out of disk space?) Look in your error log a

Re: "Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-06 Thread mos
rows into the table using Load Data Infile with the keys disabled, but over 49 hours (maybe a lot more) to rebuild 3 indexes. Even with 100g of disk space dedicated to the sort file. It started using "Repair with Filesort" for the first 7 hours, then it switched to "Repair with

Re: "Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-03 Thread Harrison Fisk
efore hand them won't help. After the data is loaded, Load Data Infile will then rebuild the keys and will choose to use "Repair with keycache" whereas "Repair with filesort" would be 100x-1000x faster. There doesn't seem to be any way to get it to use "Repa

Re: "Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-03 Thread mos
data is loaded, Load Data Infile will then rebuild the keys and will choose to use "Repair with keycache" whereas "Repair with filesort" would be 100x-1000x faster. There doesn't seem to be any way to get it to use "Repair with filesort". So I could use

Re: "Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-03 Thread Harrison Fisk
hen rebuild the keys and will choose to use "Repair with keycache" whereas "Repair with filesort" would be 100x-1000x faster. There doesn't seem to be any way to get it to use "Repair with filesort". So I could use disable keys as you had suggested and then

Re: "Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-03 Thread mos
ccurs reasonably fast that because running "Load Data Infile" on an empty table will disable the keys until all the data is loaded, so explicitly disabling before hand them won't help. After the data is loaded, Load Data Infile will then rebuild the keys and will choose to use &

Re: "Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-03 Thread Gleb Paharenko
27;t completed. A > "Show Process List" shows Info="load data infile ..." and State="Repair > with keycache". The table has a few hundred million rows of data. I assume > it is using "Repair with keycache" to rebuild the indexes after the data >

"Load Data Infile ... Repair with KeyCache" --WAY SLOW!

2005-03-02 Thread mos
I have a 50g CSV file that I am trying to import into an empty MyISAM table. It appears to go fine except after 10 hours it hasn't completed. A "Show Process List" shows Info="load data infile ..." and State="Repair with keycache". The table has a few hundred

Re: Repair with keycache

2005-02-15 Thread Kevin A. Burton
Mohamed Badri wrote: thank you for the links, I can't change system variables at runtime, so the only choice I have is, probably, to stop mysql server set variables and then run another ALTER TABLE. ;-( I'm going to siwtch to mysql4 as soon as possible. OH!... yeah... if you're on a <4 ver of

Re: Repair with keycache

2005-02-15 Thread Mohamed Badri
thank you for the links, I can't change system variables at runtime, so the only choice I have is, probably, to stop mysql server set variables and then run another ALTER TABLE. ;-( I'm going to siwtch to mysql4 as soon as possible. Selon "Kevin A. Burton" <[EMAIL PROTECTED]>: > Mohamed Badri

Re: Repair with keycache

2005-02-15 Thread Kevin A. Burton
Mohamed Badri wrote: Hi, just had a problem with a myisam table who reached 4GB of data, I increased the number of rows by doing : ALTER TABLE foo MAX_ROWS=10 This is EXACTLY what you want: http://www.peerfear.org/rss/permalink/2004/10/16/MySQLAndALTERTABLEGuiltyAsCharged http://www.peer

Repair with keycache

2005-02-15 Thread Mohamed Badri
d now the process is in state : > Repair with keycache It seems to be really slow, how can I improve speed ? or how can I get more infos about what mysql is doing ? I try to set global variables like myisam_sort_buffer_size but it didn't work. I use mysql v 3.23.57 any help wil be ap

Re: repair with keycache during ADD INDEX

2003-10-20 Thread Ken Menzel
the change then again after the change. Have you done this already? Ohterwisde your key size will be too small. Ken - Original Message - From: "Franz, Fa. PostDirekt MA" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Friday, October 17, 2003 2:34 AM Subject

repair with keycache during ADD INDEX

2003-10-16 Thread Franz, Fa. PostDirekt MA
hours, a 'SHOW PROCESSLIST' doesn't say 'copy to temp-table' anymore but 'repair with keycache'. Watching the the directory for this database, it seems to happpen at the time when all data from .MYD is copied to the temp-table. The 'repair with keycache&#

Re: Repair with keycache state while ALTERing TABLE

2002-11-23 Thread Sergei Golubchik
I have set > myisam_max_sort_file_size to 3,145,728,000 (megabytes? since I'm using > 3.23) and myisam_sort_buffer_size to 268,435,456. Yet, my database is > still using the "repair with keycache" when I add an index with "alter > table ..." Here's the r

RE: Repair with keycache state while ALTERing TABLE

2002-11-22 Thread Robert Citek
ize to 268,435,456. Yet, my database is still using the "repair with keycache" when I add an index with "alter table ..." BTW, I am using the latest stable version of MySQL (3.23.53a) on a Red Hat Linux 7.1 machine. Thanks in advance for any pointers/suggestions, - Robert -

RE: repair with keycache

2002-11-21 Thread Robert Citek
At 04:41 PM 11/21/2002 -0600, Robert Citek wrote: >So, now I know that "repair with keycache" is slower than "repair by >sorting." What can I do about that? That is, how can I make the repair >use sorting? I would have thought they would be part of the "alter

RE: repair with keycache

2002-11-21 Thread Robert Citek
the URL. >Repair with keycache - The repair code is using creating keys one by one >through the key cache. This is much slower than Repair by sorting. So, now I know that "repair with keycache" is slower than "repair by sorting." What can I do about that? That is, how can

RE: repair with keycache

2002-11-21 Thread Jennifer Goodie
http://www.mysql.com/doc/en/SHOW_PROCESSLIST.html Repair with keycache - The repair code is using creating keys one by one through the key cache. This is much slower than Repair by sorting. It is going to take a long time to build an index on a table that big, especially if you still have

repair with keycache

2002-11-21 Thread Robert Citek
Hello all, What does it mean to be in the "Repair with keycache" state, and how do I fix this? I am running MySQL v3.23.36 on a Red Hat 7.1 machine. One of my tables has about 21,000,000 (21 Million) records in it. I am trying to index one of the fields with the alter table comman

Repair with keycache state while ALTERing TABLE

2001-11-04 Thread Quentin Bennett
Hi, When doing an Alter Table, what does state Repair with keycache Mean, and should it take 27 hours (so far) on a 26,000,000 record table. And another issue: When searching the archives, I'll get something like 4. Re: Repair with keycache <http://lists.mysql.com/cgi-ez/ezmlm-c

Re: Repair with keycache

2001-07-13 Thread Sinisa Milivojevic
Mike Baranski writes: > Question about this, when I do mysqladmin processlist, I get Repair with keycache as >the state, what exactly is this? Can this be optomized? > > Mike. > > P.S. Thanks for the gcc3 patch! > You are welcome !! The above means that REPAIR ta

Repair with keycache

2001-07-13 Thread Mike Baranski
Question about this, when I do mysqladmin processlist, I get Repair with keycache as the state, what exactly is this? Can this be optomized? Mike. P.S. Thanks for the gcc3 patch! - Before posting, please check: http