mysql-20021125BK mysqld crash

Hi,
we have observer a mysqld crash. mysqld_safe did not manage to restart
mysql at all, maybe there's a bug in the restarting procedure?

How-To-Repeat:

key_buffer_size=268431360
read_buffer_size=1044480
sort_buffer_size=2097144
max_used_connections=21
max_connections=100
threads_connected=1
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_connections = 568939 K
bytes of memory
Hope that's ok; if not, decrease some variables in the equation.
[...]
0x808096b handle_segfault + 447
0x81abc38 pthread_sighandler + 176
0x81d9db3 memcpy + 51
0x8173eb0 _mi_pack_get_block_info + 316
0x8172d31 _mi_read_pack_record + 53
0x816c6e8 mi_rkey + 520
0x80d840f index_read_idx__9ha_myisamPcUiPCcUi16ha_rkey_function + 59
0x80ad082 join_read_const__FP13st_join_table + 102
0x80ace79 join_read_const_table__FP13st_join_tableP11st_position + 85
0x80a5be0 make_join_statistics__FP4JOINP13st_table_listP4ItemP16st_dyn amic_array + 2096
0x80a40b3 mysql_select__FP3THDP13st_table_listRt4List1Z4ItemP4ItemP8st _orderT4T3T4UlP13select_result + 1971
0x80a38c6 handle_select__FP3THDP6st_lexP13select_result + 102
0x808b62a mysql_execute_command__Fv + 950
0x808f255 mysql_parse__FP3THDPcUi + 149
0x808a76d dispatch_command__F19enum_server_commandP3THDPcUi + 1481
0x808a195 do_command__FP3THD + 93
0x8089a07 handle_one_connection + 635
0x81a99f8 pthread_start_thread + 216
0x81e377a thread_start + 4
[...]
New value of fp=(nil) failed sanity check, terminating stack trace!
Please read http://www.mysql.com/doc/U/s/Using_stack_trace.html and follow instructions on how to resolve the stack trace.
Resolved
stack trace is much more helpful in diagnosing the problem, so please do
resolve it
Trying to get some variables.
Some pointers may be invalid and cause the dump to abort...
thd->query at 0xb3d41b8 = select code,contig_data_code from known3d_data where id=19874
thd->thread_id=63274

I cannot say how to reproduce. I tried with most of our databases that have so many
entries, but no problem reappeared. But maybe it helps you anyway.

Linux 2.4.20-pre5-ac1 SMP -- yes, I've already prepared 2.4.20 release kernel. ;)


--
Martin Mokrejs ,
PGP5.0i key is at http://www.natur.cuni.cz/~mmokrejs
MIPS / Institute for Bioinformatics
GSF - National Research Center for Environment and Health
Ingolstaedter Landstrasse 1, D-85764 Neuherberg, Germany
tel.: +49-89-3187 3683 , fax: +49-89-3187 3585


------------------------------------------------------------ ---------
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 bugs-thread13219@lists.mysql.com
To unsubscribe, e-mail

Re: mysql-20021125BK mysqld crash

=3D?iso-8859-2?Q?Martin_MOKREJ=3DA9?=3D writes:
> Hi,
> we have observer a mysqld crash. mysqld_safe did not manage to rest=
art
> mysql at all, maybe there's a bug in the restarting procedure?
>=20
> How-To-Repeat:
>=20
> key_buffer_size=3D268431360
> read_buffer_size=3D1044480
> sort_buffer_size=3D2097144
> max_used_connections=3D21
> max_connections=3D100
> threads_connected=3D1
> It is possible that mysqld could use up to
> key_buffer_size + (read_buffer_size + sort_buffer_size)*max_connectio=
ns =3D 568939 K
> bytes of memory
> Hope that's ok; if not, decrease some variables in the equation.
> [...]
>=20
> I cannot say how to reproduce. I tried with most of our databases tha=
t have so many
> entries, but no problem reappeared. But maybe it helps you anyway.
>=20
> Linux 2.4.20-pre5-ac1 SMP -- yes, I've already prepared 2.4.20 releas=
e kernel. ;)
>=20
>=20
> --=20
> Martin Mokrejs ,
> PGP5.0i key is at http://www.natur.cuni.cz/~mmokrejs
> MIPS / Institute for Bioinformatics
> GSF - National Research Center for Environment and Health
> Ingolstaedter Landstrasse 1, D-85764 Neuherberg, Germany
> tel.: +49-89-3187 3683 , fax:=A0+49-89-3187 3585
>=20


Hi!

You know we require reproducible test cases.=20

A stack dump does not tell much , usually ....=20

--=20
Regards,
__ ___ ___ ____ __
/ |/ /_ __/ __/ __ \/ / Mr. Sinisa Milivojevic >
/ /|_/ / // /\ \/ /_/ / /__ MySQL AB, Fulltime Developer
/_/ /_/\_, /___/\___\_\___/ Larnaca, Cyprus
<___/ www.mysql.com


------------------------------------------------------------ ---------
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 bugs-thread13220@lists.mysql.com
To unsubscribe, e-mail

Re: mysql-20021125BK mysqld crash

Hi,

Same problem for me :

mysql> EXPLAIN SELECT 1 FROM (SELECT 1) a PROCEDURE ANALYSE();
ERROR 2013: Lost connection to MySQL server during query
mysql>
mysql> \s
--------------
mysql Ver 13.0 Distrib 4.1.0-alpha, for pc-linux (i686)

Connection id: 67101
SSL: Not in use
Current pager: stdout
Using outfile: ''
Server version: 4.1.0-alpha
Protocol version: 10
Connection: Localhost via UNIX socket
Client characterset: latin1
Server characterset: latin1
UNIX socket: /var/adm/mysql/mysql.sock
--------------

mysql> SELECT 1;
ERROR 2006: MySQL server has gone away
No connection. Trying to reconnect...
ERROR 1049: Unknown database '-­Hardwarefr'
ERROR:
Can't connect to the server

mysql> use Hardwarefr
Database changed

mysql> SELECT 1;
No connection. Trying to reconnect...
ERROR 1049: Unknown database 'yHardwarefr'
ERROR:
Can't connect to the server

It seems the database name is corrupted ?!

Regards,
Jocelyn

----- Original Message -----
From: "Martin MOKREJ©"
To:
Sent: Thursday, December 05, 2002 1:31 PM
Subject: mysql-20021125BK mysqld crash


> Hi,
> we have observer a mysqld crash. mysqld_safe did not manage to restart
> mysql at all, maybe there's a bug in the restarting procedure?


------------------------------------------------------------ ---------
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 bugs-thread13222@lists.mysql.com
To unsubscribe, e-mail

Re: mysql-20021125BK mysqld crash

On Thu, 5 Dec 2002, Sinisa Milivojevic wrote:

> > I cannot say how to reproduce. I tried with most of our databases that have so many
> > entries, but no problem reappeared. But maybe it helps you anyway.

> You know we require reproducible test cases.
>
> A stack dump does not tell much , usually ....

Yes, I know, but sometimes you can figure out the problem anyway. ;))
And yes, the mysqld_safe problem might be reproducible on your side too.
;) It is just a trial to point you to some part of the code, nothing more.


--
Martin Mokrejs ,
PGP5.0i key is at http://www.natur.cuni.cz/~mmokrejs
MIPS / Institute for Bioinformatics
GSF - National Research Center for Environment and Health
Ingolstaedter Landstrasse 1, D-85764 Neuherberg, Germany
tel.: +49-89-3187 3683 , fax: +49-89-3187 3585


------------------------------------------------------------ ---------
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 bugs-thread13224@lists.mysql.com
To unsubscribe, e-mail

Re: mysql-20021125BK mysqld crash

=3D?iso-8859-2?Q?Martin_MOKREJ=3DA9?=3D writes:
> On Thu, 5 Dec 2002, Sinisa Milivojevic wrote:
>=20
>=20
> Yes, I know, but sometimes you can figure out the problem anyway. ;))=

> And yes, the mysqld_safe problem might be reproducible on your side t=
oo.
> ;) It is just a trial to point you to some part of the code, nothing =
more.
>=20
>=20
> --=20
> Martin Mokrejs ,
> PGP5.0i key is at http://www.natur.cuni.cz/~mmokrejs
> MIPS / Institute for Bioinformatics
> GSF - National Research Center for Environment and Health
> Ingolstaedter Landstrasse 1, D-85764 Neuherberg, Germany
> tel.: +49-89-3187 3683 , fax:=A0+49-89-3187 3585
>=20
>=20

Of course, but the above is true if we have solved this bug alrady and
we recognize symptoms from the stack trace, which we do not from your
stack trace. Which means you could have hit a new bug .... ;-(

--=20
Regards,
__ ___ ___ ____ __
/ |/ /_ __/ __/ __ \/ / Mr. Sinisa Milivojevic >
/ /|_/ / // /\ \/ /_/ / /__ MySQL AB, Fulltime Developer
/_/ /_/\_, /___/\___\_\___/ Larnaca, Cyprus
<___/ www.mysql.com


------------------------------------------------------------ ---------
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 bugs-thread13226@lists.mysql.com
To unsubscribe, e-mail

mysql-20021125BK mysqld crash

Hi!

>>>>> "Martin" == Martin MOKREJ writes:

Martin> Hi,
Martin> we have observer a mysqld crash. mysqld_safe did not manage to restart
Martin> mysql at all, maybe there's a bug in the restarting procedure?

Martin> How-To-Repeat:



Martin> 0x808096b handle_segfault + 447
Martin> 0x81abc38 pthread_sighandler + 176
Martin> 0x81d9db3 memcpy + 51
Martin> 0x8173eb0 _mi_pack_get_block_info + 316
Martin> 0x8172d31 _mi_read_pack_record + 53

The above indicates that the table was somehow corrupted.
A REPAIR TABLE should fix this.



Martin> I cannot say how to reproduce. I tried with most of our databases that have so many
Martin> entries, but no problem reappeared. But maybe it helps you anyway.

Sorry, but I don't know why MySQL didn't restart; One possibility is
a bug in Linuxthreads so that the process didn't terminate properly
and thus the mysqld_safe script didn't notice that mysqld died.

Maybe you can find more about this in the mysql.err file...


Regards,
Monty

--
For technical support contracts, goto https://order.mysql.com/
__ ___ ___ ____ __
/ |/ /_ __/ __/ __ \/ / Mr. Michael Widenius
/ /|_/ / // /\ \/ /_/ / /__ MySQL AB, CTO
/_/ /_/\_, /___/\___\_\___/ Helsinki, Finland
<___/ www.mysql.com

------------------------------------------------------------ ---------
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 bugs-thread13254@lists.mysql.com
To unsubscribe, e-mail

Re: mysql-20021125BK mysqld crash

On Tue, 10 Dec 2002, Michael Widenius wrote:

Hi,

> >>>>> "Martin" == Martin MOKREJ writes:
>
> Martin> Hi,
> Martin> we have observer a mysqld crash. mysqld_safe did not manage to restart
> Martin> mysql at all, maybe there's a bug in the restarting procedure?
>
> Martin> How-To-Repeat:
>
>
>
> Martin> 0x808096b handle_segfault + 447
> Martin> 0x81abc38 pthread_sighandler + 176
> Martin> 0x81d9db3 memcpy + 51
> Martin> 0x8173eb0 _mi_pack_get_block_info + 316
> Martin> 0x8172d31 _mi_read_pack_record + 53
>
> The above indicates that the table was somehow corrupted.
> A REPAIR TABLE should fix this.

Michael, I've found this night mysqld (same installation), crashed on the
same machine with exactly same trace. The select command was same
(operating on same table, selecting same 2 columns, only the row was
different).

I've checked all tables of a given name on the machine, none of then is
marked as crashed. I did not use 'myisamchk -dvv' but CHECK TABLE.

Are you sure the table was crashed? I'm going to run infinite loop repeating this select command.
My question is, was that table in cache or not? Or, does it matter?

>
>
>
> Martin> I cannot say how to reproduce. I tried with most of our databases that have so many
> Martin> entries, but no problem reappeared. But maybe it helps you anyway.
>
> Sorry, but I don't know why MySQL didn't restart; One possibility is
> a bug in Linuxthreads so that the process didn't terminate properly
> and thus the mysqld_safe script didn't notice that mysqld died.
>
> Maybe you can find more about this in the mysql.err file...

No, nothing:

021210 22:44:10 mysqld ended

021211 09:27:09 mysqld started
021211 9:27:09 /usr/local/mysql/libexec/mysqld: Can't create/write to file '/usr/local/mysql/run/jerboas.pid' (Errcode: 13)
/usr/local/mysql/libexec/mysqld: ready for connections

The permissions problem I've corrected.

--
Martin Mokrejs ,
PGP5.0i key is at http://www.natur.cuni.cz/~mmokrejs
MIPS / Institute for Bioinformatics
GSF - National Research Center for Environment and Health
Ingolstaedter Landstrasse 1, D-85764 Neuherberg, Germany
tel.: +49-89-3187 3683 , fax: +49-89-3187 3585


------------------------------------------------------------ ---------
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 bugs-thread13258@lists.mysql.com
To unsubscribe, e-mail

Re: mysql-20021125BK mysqld crash

=3D?iso-8859-2?Q?Martin_MOKREJ=3DA9?=3D writes:
> On Tue, 10 Dec 2002, Michael Widenius wrote:
>=20
> Hi,
>=20
> 021210 22:44:10 mysqld ended
>=20
> 021211 09:27:09 mysqld started
> 021211 9:27:09 /usr/local/mysql/libexec/mysqld: Can't create/write =
to file '/usr/local/mysql/run/jerboas.pid' (Errcode: 13)
> /usr/local/mysql/libexec/mysqld: ready for connections
>=20
> The permissions problem I've corrected.
>=20
> --=20
> Martin Mokrejs ,
> PGP5.0i key is at http://www.natur.cuni.cz/~mmokrejs
> MIPS / Institute for Bioinformatics
> GSF - National Research Center for Environment and Health
> Ingolstaedter Landstrasse 1, D-85764 Neuherberg, Germany
> tel.: +49-89-3187 3683 , fax:=A0+49-89-3187 3585
>=20

This bug is not caused by permissions problem, but others could have
been, like no core files or problem in re-starting ...

--=20
MySQL 2003 Users Conference -> http://www.mysql.com/events/uc2003/

__ ___ ___ ____ __
/ |/ /_ __/ __/ __ \/ / Mr. Sinisa Milivojevic >
/ /|_/ / // /\ \/ /_/ / /__ MySQL AB, Fulltime Developer
/_/ /_/\_, /___/\___\_\___/ Larnaca, Cyprus
<___/ www.mysql.com


------------------------------------------------------------ ---------
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 bugs-thread13260@lists.mysql.com
To unsubscribe, e-mail

Re: mysql-20021125BK mysqld crash

On Wed, 11 Dec 2002, Sinisa Milivojevic wrote:

> =?iso-8859-2?Q?Martin_MOKREJ=A9?= writes:
> > On Tue, 10 Dec 2002, Michael Widenius wrote:
> >
> > Hi,
> >
> > 021210 22:44:10 mysqld ended
> >
> > 021211 09:27:09 mysqld started
> > 021211 9:27:09 /usr/local/mysql/libexec/mysqld: Can't create/write to file '/usr/local/mysql/run/jerboas.pid' (Errcode: 13)
> > /usr/local/mysql/libexec/mysqld: ready for connections
> >
> > The permissions problem I've corrected.
>
> This bug is not caused by permissions problem, but others could have
> been, like no core files or problem in re-starting ...

Are you sure? Only the PID file is create dthere, logfile and socketfile
are in another directory. I don't want to belive that it could be related
to problem will dumping a core. At how many places is mysql operating on a
pid file? Are there checks that it succeeded opening/writing/closing the
file? If you really think so, would it be possible to force mysql to exit,
when it cannot PID file on the startup? Then it will force the admins to
fix the directory permissions immediately. ;)

Regarding repeating the problem. I 've 6 databases which have so many rows
in that table. In total, I've so far 1178025 select, and no crash yet.
Probably I should flush the cache always before.

If Michael thinks the table was broken, then when I hit the cache, it
probably won't manifest, right?

--
Martin Mokrejs ,
PGP5.0i key is at http://www.natur.cuni.cz/~mmokrejs
MIPS / Institute for Bioinformatics
GSF - National Research Center for Environment and Health
Ingolstaedter Landstrasse 1, D-85764 Neuherberg, Germany
tel.: +49-89-3187 3683 , fax: +49-89-3187 3585


------------------------------------------------------------ ---------
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 bugs-thread13261@lists.mysql.com
To unsubscribe, e-mail