Bug #6781 mySQL crashes from executing init_file
Submitted: 23 Nov 2004 21:02 Modified: 19 Jan 2005 11:36
Reporter: Dathan Pattishall Email Updates:
Status: No Feedback Impact on me:
None 
Category:MySQL Server Severity:S1 (Critical)
Version:4.0.20 OS:Linux (Linux dbcountmstr1 2.4.21-215-sm)
Assigned to: CPU Architecture:Any

[23 Nov 2004 21:02] Dathan Pattishall
Description:
mySQL-4.0.20 crashes on startup when the following events.

Use this my.cnf

# The following options will be passed to all MySQL clients
# $Id: my.cnf.tmpl,v 1.1 2004/11/22 23:06:21 dathan Exp $
[client]
port>--->---= 3306
socket>->---= /var/lib/mysql/mysql.sock

# Here follows entries for some specific programs

# The MySQL server
[mysqld]
port            = 3306
socket>->---= /var/lib/mysql/mysql.sock
skip-external-locking
skip-name-resolve
skip-innodb
skip-bdb
key_buffer=1024M
max_allowed_packet=16M
table_cache=512
tmp_table_size  = 64M
sort_buffer_size=4M
read_buffer_size=4M
read_rnd_buffer_size=4M
thread_cache=20
myisam_sort_buffer_size=32M
query_cache_type = 1
query_cache_size = 64M
#init_file crashes mysql at startup
#init_file=dbcntinit.sql

# Replication settings (must be unique across master and all slaves)
server-id>--= %SERVERID%

# Table Filters (only replicate these tables - only used for slow-read slaves)
slave-skip-errors=1062

# Create slow query log (WARNING: can consume a lot of disk on slow-read slaves)
long_query_time=10
max_connections=300
max_connect_errors=999999
wait_timeout=120

# Point the following paths to different dedicated disks
tmpdir>->---    = /tmp/>>---
relay-log >-    = /var/lib/mysql/dbslave-relay-bin
master-info-file    = /var/lib/mysql/master.info
relay-log-index>    = /var/lib/mysql/dbslave-relay-bin.index
relay-log-info-file = /var/lib/mysql/relay-log.info
log-slave-updates
log-bin             = /var/lib/mysql/dbcnt-bin

[mysqldump]
quick
max_allowed_packet=4M

[mysql]
no-auto-rehash
# Remove the next comment character if you are not familiar with SQL
#safe-updates

[isamchk]
key_buffer=256M
sort_buffer=256M
read_buffer=2M
write_buffer=2M

[myisamchk]
key_buffer=256M
sort_buffer=256M
read_buffer=2M
write_buffer=2M

[mysqlhotcopy]
interactive-timeout

slave stop;
CREATE DATABASE IF NOT EXISTS clickcounts;
use clickcounts;
CREATE TABLE IF NOT EXISTS profile_view_count_sum
(
 user_id int NOT NULL PRIMARY KEY,
 count_total smallint  unsigned NOT NULL DEFAULT 0,
 count_since date default NULL
)  TYPE=MYISAM DATA DIRECTORY = '/dev/shm/clickcounts' INDEX DIRECTORY = '/dev/s
hm/clickcounts';

--
-- 4 + 4 + 4 = 16 bytes per row: 20 million users is 228 MB
--
CREATE TABLE IF NOT EXISTS profile_view_count_detail
(
 viewed_id int not null,
 viewer_id int not null,
 viewdate  date default null,
 PRIMARY KEY (viewed_id,viewer_id,viewdate)
) TYPE=MYISAM DATA DIRECTORY = '/dev/shm/clickcounts' INDEX DIRECTORY = '/dev/sh
m/clickcounts';

slave start

The slave is not configured

How to repeat:
Look above

Suggested fix:
look at the init_file code in src/sql
[7 Dec 2004 19:43] Victoria Reznichenko
Hi,

Thank you for the report. I tested your example with 4.0.23-debug-log and MySQL server started successfully with init_file.
Does MySQL server crash if you comment init_file option in the my.cnf?
[7 Dec 2004 20:10] Dathan Pattishall
The server does not crash if I comment out the init_file field  in my.cnf

In fact I had to resort to starting mysqld then in the init script process use the client to apply the changes of the init_file sql file.

i.e.

mysql -uroot  mydatabase < init_file.sql
[19 Dec 2004 11:36] Hartmut Holzgraefe
Thank you for this bug report. To properly diagnose the problem, we
need a backtrace to see what is happening behind the scenes. To
find out how to generate a backtrace, please read
http://www.mysql.com/doc/en/Making_trace_files.html

Once you have generated a backtrace, please submit it to this bug
report and change the status back to 'Open'. Thank you for helping
us make our products better.
[14 Feb 2005 22:54] Bugs System
No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".