| Bug #57547 | mysqlslap: Cannot drop database 'mysqlslap' ERROR : MySQL server has gone away | ||
|---|---|---|---|
| Submitted: | 19 Oct 2010 4:45 | Modified: | 21 Nov 2010 19:18 |
| Reporter: | Reza Iskandar Achmad | Email Updates: | |
| Status: | No Feedback | Impact on me: | |
| Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S3 (Non-critical) |
| Version: | mysql-5.1-telco-7.1 | OS: | Linux |
| Assigned to: | CPU Architecture: | Any | |
| Tags: | 5.1.47-ndb-7.1.8 | ||
[19 Oct 2010 4:45]
Reza Iskandar Achmad
Error log from management node
Attachment: ndb_mgm.txt (text/plain), 3.08 KiB.
[19 Oct 2010 4:50]
Reza Iskandar Achmad
Cluster config.ini
Attachment: config.ini (application/octet-stream, text), 3.45 KiB.
[19 Oct 2010 4:52]
Reza Iskandar Achmad
SQL Node config
Attachment: my.cnf (application/octet-stream, text), 1.81 KiB.
[19 Oct 2010 7:34]
Magnus Blåudd
Looks like you manage to crash the mysqld because of overload. Could be the "Event buffer" or something else. We need more log files if possible. Maybe you can setup mysqld to generate a core when it crashes. There should also be a log file from mysqld. > 2010-10-19 11:35:59 [MgmtSrvr] INFO -- Node 21: Event buffer status: used=2133KB(100%) alloc=2133KB(0%) max=0B apply_epoch=301065/7 latest_epoch=301101/1 > 2010-10-19 11:35:59 [MgmtSrvr] INFO -- Node 21: Event buffer status: used=23KB(1%) alloc=2133KB(0%) max=0B apply_epoch=301101/2 latest_epoch=301101/2 I agree these lines look suspicious, but they don't give the full picture which the mysqld log file would.
[19 Oct 2010 8:04]
Reza Iskandar Achmad
so I started mysqld like this: mysqld --basedir=/srv/mysql/ --datadir=/srv/mysql/data --core --log-error=/srv/mysql/data/ndb-mysqld-01.err --open-files-limit=65535 --pid-file=/srv/mysql/data/mysqld.pid --socket=/srv/mysql/data/mysql.sock --port=3306 but could not find the core file. Could it be mysqld not crashed?
[19 Oct 2010 8:25]
Reza Iskandar Achmad
I believe this is not mysqld problem. I just tested to change mysqld configuration to not use ndbcluster and MyISAM as default storage engine: $ mysqlslap -uyyyyy -pxxxxxx -h 172.21.0.101 --auto-generate-sql --auto-generate-sql-add-autoincrement -c 3500 Benchmark Average number of seconds to run all queries: 5.290 seconds Minimum number of seconds to run all queries: 5.290 seconds Maximum number of seconds to run all queries: 5.290 seconds Number of clients running queries: 3500 Average number of queries per client: 0
[21 Oct 2010 19:18]
Sveta Smirnova
Thank you for the feedback. Please send MySQL error log file also.
[22 Nov 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".
Description: We keep to receive this message when testing the cluster. We are using Ubuntu Lucid standard kernel. mysql@ndb-mysqld-01:~$ ulimit -a | grep "open files" open files (-n) 132000 it only happens when testing database using ndbcluster engine How to repeat: 1. Run Cluster: 2 mgm node; 2 SQL Node; 2 Data node 2. Run mysql slap from one of SQL node $ mysqlslap -u root -p -S data/mysql.sock --auto-generate-sql --auto-generate-sql-add-autoincrement -c 2000 3. Management node gives warning about event buffer and ended with mysql server gone away