Bug #102 Test failure in test "rpl000015"
Submitted: 26 Feb 2003 6:48 Modified: 28 Feb 2003 2:30
Reporter: Lenz Grimmer Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Server: Replication Severity:S2 (Serious)
Version:4.1.0 OS:Linux (Linux/alpha)
Assigned to: Guilhem Bichot CPU Architecture:Any

[26 Feb 2003 6:48] Lenz Grimmer
Description:
On our Linux/alpha build host, the current 4.1 tree fails with the following error: 
 
rpl000013                     0.01     0.01     8.05       [ pass ] 
rpl000015                     ....     ....     ....       [ fail ] 
 
   >> Error at line 36: Result length mismatch<< 
Below are the diffs between actual and expected results: 
------------------------------------------------------- 
*** r/rpl000015.result  Wed Feb 26 15:16:15 2003 
--- r/rpl000015.reject  Wed Feb 26 17:06:03 2003 
*************** 
*** 5,23 **** 
  reset slave; 
  show slave status; 
  Master_Host   Master_User     Master_Port     Connect_retry   Master_Log_File 
  change master to master_host='127.0.0.1'; 
  show slave status; 
  Master_Host   Master_User     Master_Port     Connect_retry   Master_Log_File 
! 127.0.0.1     test    MASTER_PORT     7               4       slave-relay-bin. 
  change master to master_host='127.0.0.1',master_user='root', 
  master_password='',master_port=MASTER_PORT; 
  show slave status; 
  Master_Host   Master_User     Master_Port     Connect_retry   Master_Log_File 
! 127.0.0.1     root    MASTER_PORT     7               4       slave-relay-bin. 
  start slave; 
  show slave status; 
  Master_Host   Master_User     Master_Port     Connect_retry   Master_Log_File 
! 127.0.0.1     root    MASTER_PORT     7       master-bin.000001       79 
  drop table if exists t1; 
  create table t1 (n int); 
  insert into t1 values (10),(45),(90); 
--- 5,24 ---- 
  reset slave; 
  show slave status; 
  Master_Host   Master_User     Master_Port     Connect_retry   Master_Log_File 
+ 127.0.0.1     root    MASTER_PORT     1               4       slave-relay-bin. 
  change master to master_host='127.0.0.1'; 
  show slave status; 
  Master_Host   Master_User     Master_Port     Connect_retry   Master_Log_File 
! 127.0.0.1     root    MASTER_PORT     1               4       slave-relay-bin. 
  change master to master_host='127.0.0.1',master_user='root', 
  master_password='',master_port=MASTER_PORT; 
  show slave status; 
  Master_Host   Master_User     Master_Port     Connect_retry   Master_Log_File 
! 127.0.0.1     root    MASTER_PORT     1               4       slave-relay-bin. 
  start slave; 
  show slave status; 
  Master_Host   Master_User     Master_Port     Connect_retry   Master_Log_File 
! 127.0.0.1     root    MASTER_PORT     1       master-bin.000001       79 
  drop table if exists t1; 
  create table t1 (n int); 
  insert into t1 values (10),(45),(90); 
------------------------------------------------------- 
 

How to repeat:
[26 Feb 2003 15:04] Guilhem Bichot
Problem reproduced on build.mysql2.com.
mysql-test-run --local t/rpl000015.test

Reason is this host probably pulled badly the latest changes.
Yesterday Monty did some renaming on
mysql-test/t/rpl000015-slave-master-info.opt
mysql-test/t/rpl_rotate_logs-slave-master-info.opt
and updated mysql-test-run.sh as required
(this was to fix a rpl000015 problem on MAC OS X where filenames
can't be too long).
Then auto-merge 4.0->4.1 occured on 26 Feb 00:54
But it seems that the 4.1 tree on build.mysql2.com didn't catch it
all fine, because it has the modified mysql-test-run but the
files haven't be renamed.
While I did a 4.1 pull on my machine right now and I have both changes
(they're in changeset 1.676.1.661).
So the problem is a bad pull or something like this.
[27 Feb 2003 6:11] Lenz Grimmer
I tried it today with a fresh pull - still the same problem. It fails on 
Linux (s390/ia64/x86/alpha) and Mac OS X 10.2.4, HPUX11 (32bit). 
 
Last ChangeSet was: 1.1478 
 
Could it be the compile options used? I use the one for the "standard" binary 
(Linux/x86): 
 
CFLAGS="-O2 -mcpu=pentiumpro" CXX=gcc CXXFLAGS="-O2 
-mcpu=pentiumpro -felide-constructors" ./configure 
--prefix=/usr/local/mysql --with-comment="Official MySQL-standard 
binary" --with-extra-charsets=complex --with-server-suffix="-standard" 
--enable-thread-safe-client --enable-local-infile --enable-assembler 
--disable-shared --with-client-ldflags=-all-static 
--with-mysqld-ldflags=-all-static --with-innodb
[28 Feb 2003 2:30] Michael Widenius
Thank you for your bug report. This issue has been fixed in the latest
development tree for that product. You can find more information about
accessing our development trees at 
    http://www.mysql.com/doc/en/Installing_source_tree.html

The problem here is that there is missing files in the 4.1 distribution:

Here is the file on my development machine:

((((/my/mysql-4.1/mysql-test)))) ls t/rpl000015*
t/rpl000015-slave.sh  t/rpl000015.slave-mi  t/rpl000015.test

On ds20:

mysqldev@ds20:~/ds20/mysql-4.1.0-alpha/mysql-test > ls t/rpl000015*
t/rpl000015-slave.sh  t/rpl000015.test

The problem was that we did not have a rule for *.slave-min in mysql-test/Makefile.am