Project

General

Profile

Actions

Bug #2942

closed

many access at the same time seems to create errors

Added by Dosme Nicolas almost 12 years ago. Updated almost 11 years ago.

Status:
Closed
Priority:
Low
Assigned To:
Category:
-
Start date:
07/16/2012
Due date:
05/12/2013
% Done:

0%

Estimated time:

Description

When GEC have lots of request at the same time to the Topology Manager there is an error message on the TM side:

.....
 10 INFO [main_task_00000000018D7A30] Topology Manager  - Topology Manager initialized !
 10 INFO [main_task_00000000018D7A30] TM XML access - Logger launched !
 10 INFO [main_task_00000000018D7A30] TM XML access - Logger launched !
 123 INFO [A778b(5)_0000000001C50A40] TM XML access - Topology_manager->Get_List_Db_Items
 123 INFO [A778b(5)_0000000001C50A40] TM XML access - Topology_manager->Get_List_Db_Items
Error in my_thread_global_end(): 2 threads didn't exit
Error in my_thread_global_end(): 5 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
Error in my_thread_global_end(): 8 threads didn't exit
....

when this error message appears the TM is very slow to answer...

You can reproduce this "feature" :) with this script

#!/bin/sh

EXEC=/tools/TM/bin/tm_xml_access
FUNC=get_information

for item in ATC04 ATC05 A005 A007 A010 A008 A009 C001 ADC01 C002 ADC02 C003 ADC03 C004 ADC04 C005 ADC05 C006 C007 C010 C008 C009 B001 B002 B003 B004 B005 B006 B007 B010 B008 B009 ATC01 ATC02 ATC03
do
($EXEC $FUNC $item&)
($EXEC $FUNC $item&)
($EXEC $FUNC $item&)
($EXEC $FUNC $item&)
($EXEC $FUNC $item&)
($EXEC $FUNC $item&)
($EXEC $FUNC $item&)
($EXEC $FUNC $item&)
($EXEC $FUNC $item&)
done


Files

Actions #1

Updated by Guevara Luz almost 12 years ago

  • Status changed from New to Feedback

This bug is reported in the debian page as : #680412.
Error in my_thread_global_end: x threads didn't exit (MySQL bug #57727)
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=680412

We are using Debian GNU/Linux 6.0.2 squeeze (stable). The bug page says 'the bug was fixed in version 5.1.8 of package libmyodbc'. We (Vincent and me) have tried to migrate from libmyodbc-5.1.6-1 to libmyodbc-5.1.8 as well as the whole libraries that it depends on (26 libraries).

Following the compilation of the topology manager with these new library versions and after resolving all path problems, we obtain the followings errors:

_Topology_Manager : ErrorGNU.DB.SQLCLI.CONNECTION_ERROR [Proc=SQLConnect][Server=][State=08S01][unixODBC][MySQL][ODBC 5.1 Driver]Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2)

0 ERROR [main_task_000000000175AA30] Topology Manager - Topology_Manager Error : GNU.DB.SQLCLI.CONNECTION_ERROR

0 ERROR [main_task_000000000175AA30] Topology Manager - [Proc=SQLConnect][Server=][State=08S01][unixODBC][MySQL][ODBC 5.1 Driver]Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2)
_

I did not recompile the gnade packages against the latest libraries version and perhaps this could explain the origin of these problems.
I think the best solution is to recompile the gnade packages and test the topology manager inside a wheezy (testing) environment.

Any new ideas are welcome.

Actions #2

Updated by Dosme Nicolas almost 12 years ago

We are testing the TM at GSI and we got that:

 1 INFO [main_task_0000000002045890] Topology Manager  - Activable_Container : Adding device name << mainboard0030 >> kind << MAINBOARD >> to device tree of << Digitizer030 >>
 1 INFO [main_task_0000000002045890] Topology Manager  - Activable_Container : Adding device name << carrier031 >> kind << CARRIER >> to device tree of << C010 >>
 1 INFO [main_task_0000000002045890] Topology Manager  - Activable_Container : Adding device name << carrier033 >> kind << CARRIER >> to device tree of << C010 >>
 1 INFO [main_task_0000000002045890] Topology Manager  - Activable_Container : Adding device name << anode30 >> kind << NARVAL_NODE >> to device tree of << C010 >>
 1 INFO [main_task_0000000002045890] Topology Manager  - Activable_Container : Adding device name << gts054 >> kind << GTS >> to device tree of << C010 >>
 1 WARN [main_task_0000000002045890] Topology Manager  -  
 1 INFO [main_task_0000000002045890] Topology Manager  - Topology Manager initialized !
 1 INFO [main_task_0000000002045890] TM XML access - Logger launched !
 1 INFO [main_task_0000000002045890] TM XML access - Logger launched !
 13 INFO [A778b(1)_00000000021585C0] TM XML access - Topology_manager->Get_List_Db_Items
 13 INFO [A778b(1)_00000000021585C0] TM XML access - Topology_manager->Get_List_Db_Items
Error in my_thread_global_end(): 1 threads didn't exit
tm_server: pthread_mutex_lock.c:62: __pthread_mutex_lock: Assertion `mutex->__data.__owner == 0' failed.
Aborted

and we got back the prompt... :( it is not only slow, it crash now... :(

Actions #3

Updated by Guevara Luz almost 12 years ago

What is your TM version?
Could you attach your configurations files: db_item, activation_model and hadware_tree XML files, please?

Actions #4

Updated by Dosme Nicolas almost 12 years ago

version is r125 we switch to the "cluster version".

db_item, activation_model and database.sql are provided by the svn... r125.

the hadware_tree XML file use is the one attached.

Actions #5

Updated by Grave Xavier almost 11 years ago

  • Due date set to 05/12/2013
  • Status changed from Feedback to Closed
  • Assigned To set to Grave Xavier

Test with the script done with the new TM : no error, no slowing down

Actions

Also available in: Atom PDF