Bug #27510 Got error 127 when reading table, but no crashed table
Submitted: 29 Mar 2007 7:12 Modified: 9 Jan 2010 17:54
Reporter: César Alcaide Email Updates:
Status: No Feedback Impact on me:
None 
Category:MySQL Server Severity:S3 (Non-critical)
Version:5.0.37, 5.0.36 OS:Linux (Linux kernel 2.4.20)
Assigned to: CPU Architecture:Any
Tags: corruption, error 127

[29 Mar 2007 7:12] César Alcaide
Description:
mysqld logs error:
[ERROR] Got error 127 when reading table './Aris/puerto'
but table is not crashed

It occurs when trying to read a row recently inserted

show create table puerto:
CREATE TABLE `puerto` (
  `auto_n` double NOT NULL auto_increment,
  `texto` text,
  `Int_MD` timestamp NOT NULL default CURRENT_TIMESTAMP on update CURRENT_TIMESTAMP,
  PRIMARY KEY  (`auto_n`)
) ENGINE=MyISAM AUTO_INCREMENT=8656 DEFAULT CHARSET=latin1 

How to repeat:
ejex("truncate table puerto");
ejex("insert into puerto (texto) values ('testing')");

while(true)
 $a = tot("select texto from puerto where auto_n = 1");

Suggested fix:
Perhaps there is an litle interval of time when mysql believes record exists, but not yet available.
[29 Mar 2007 11:06] Valeriy Kravchuk
Thank you for a problem report. Have you tried to run CHECK TABLE for that table? Can you upload a dump of it?
[29 Mar 2007 19:08] César Alcaide
check table reports OK. The problem occurs even creating the table before.
[8 Apr 2007 10:24] Lanny Walker
I just received the same error.

070408  9:59:36 [ERROR] Got error 127 when reading table './buyer/item'

mysql> check table item;
+------------+-------+----------+----------+
| Table      | Op    | Msg_type | Msg_text |
+------------+-------+----------+----------+
| buyer.item | check | status   | OK       |
+------------+-------+----------+----------+
1 row in set (1 min 33.55 sec)

MySQL-server-enterprise-gpl-5.0.36-0.rhel4

Linux hostname.obscured.com 2.6.9-42.ELsmp #1 SMP Tue Aug 15 10:35:26 BST 2006 x86_64 x86_64 x86_64 GNU/Linux
[12 Apr 2007 8:48] Valeriy Kravchuk
All reporters:

Pleased, send the results of:

getconf GNU_LIBC_VERSION
getconf GNU_LIBPTHREAD_VERSION
[12 Apr 2007 9:54] César Alcaide
getconf GNU_LIBC_VERSION
glibc 2.3.2
[root@infrarrlinuxaris root]# getconf GNU_LIBPTHREAD_VERSION
NPTL 0.34
[12 Apr 2007 10:15] Valeriy Kravchuk
César,

Please, upgrade your glibc! With that:

glibc 2.3.2
NPTL 0.34

you are doomed. Alternatively, add

export LD_ASSUME_KERNEL=2.4.1

to the script used to start MySQL server, restart it, and check if it will influence your case.
[12 May 2007 23:00] 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".
[24 Jun 2007 23:00] 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".
[9 Dec 2009 17:47] Sheeri Cabral
I am getting this same problem, 

$ getconf GNU_LIBC_VERSION
glibc 2.7
$ getconf GNU_LIBC_VERSION
glibc 2.7

$ uname -a
Linux [machinename] 2.6.26.3-14.fc8 #1 SMP Wed Sep 3 03:17:52 EDT 2008 x86_64 x86_64 x86_64 GNU/Linux

$ cat /etc/issue
Fedora release 8 (Werewolf)
Kernel \r on an \m
[9 Dec 2009 17:54] Valeriy Kravchuk
Sheeri,

What version of MySQL server do you use? Send also the results of SHOW CREATE TABLE and SHOW TABLE STATUS for the problematic table.
[9 Dec 2009 21:57] Sheeri Cabral
Actually, I believe it's related to this bug:  

http://bugs.mysql.com/bug.php?id=29838

Which I'm surprised isn't linked here at all.  The server is MySQL 5.0.45, hence why I think it's bug 29838, it was a bug fixed in MySQL 5.0.51.  I will upgrade and see if it's still a problem.

You should link the bugs, though, if they are related.
[11 Dec 2009 8:20] Sveta Smirnova
Sheeri,

thank you for the update. Bugs are linked. We will wait results of your testing.
[10 Jan 2010 0:00] 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".