Bug #11015 massive concurrent inserts cause error or server crash
Submitted: 1 Jun 2005 12:03 Modified: 3 Jul 2005 11:31
Reporter: Pierre COLOMBIER Email Updates:
Status: No Feedback Impact on me:
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S1 (Critical)
Version:4.1.11 OS:Linux (Linux)
Assigned to: CPU Architecture:Any

[1 Jun 2005 12:03] Pierre COLOMBIER
when i insert a lot of data in the cluster using the mysql client of an API node, there is no problem. If I make multiple operations like that, I get some errors about duplicate entries

ERROR 1062 at line 7843: Duplicate entry '' for key 65

sometime it gives 

ERROR 1205 at line 141824: Lock wait timeout exceeded; try restarting transaction

and the whole cluster crash

How to repeat:
cluster with 1 admin node
2 api nodes
10 storage nodes.

on api node 1 or 2
lots of insertions with command

cat file1.sql | mysql database_name &
cat file2.sql | mysql database_name &
cat file3.sql | mysql database_name &

ERROR 1062 at line [random_line]: Duplicate entry '' for key 65

(I'm absolutely sure file1.sql, file2.sql and file3.sql 
don't have any duplicate entries.

same process but without parallelism 

cat file1.sql file2.sql file3.sql  | mysql database_name

don't cause this error.


sometimes, the whole cluster crash.
[3 Jun 2005 11:31] Martin Skold
Please attach error and trace files for crash.
See Reference Manual, 
chapter16.5.2. ndbd, the Storage Engine Node Process
[3 Jul 2005 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".