Hello,
A month ago, I have used innobackupex and this protocol
[URL=“How to setup a slave for replication in 6 simple steps with Percona XtraBackup”]http://www.percona.com/doc/percona-x...plication.html[/URL]
to set up two slaves.
Today I tried to reset one of the slaves from scratch using exactly the same protocol. The innobackupex program took considerable time (the program was printing something like “log position …”. Meanwile the server started to run out of available connections. I hade to kill innobackupex and stop/start mysql when there were 500 waiting connections in mysql.
I bellieve that innobackupex should lock/unlock the databases one by one, but the waiting connections were from many different databases. This is a production server with SSD disks and enough RAM.
mysql> show variables like “%version%”;
±------------------------±---------------------+
| Variable_name | Value |
±------------------------±---------------------+
| innodb_version | 5.5.37 |
| protocol_version | 10 |
| slave_type_conversions | |
| version | 5.5.37-0+wheezy1-log |
| version_comment | (Debian) |
| version_compile_machine | x86_64 |
| version_compile_os | debian-linux-gnu |
±------------------------±---------------------+
EDIT:
I am not sure if this was a locking problem. What I know for sure is:
- after starting innobackupex the number of connections in show_process_list increased from standard 10-20 to 512
- afreaching 500 connections, I killed innobackupex, but the number of connections did not decrease - still 512connections
- after several minutes I decided to stop/start the mysql server
this is the reason why I suppose that innodbbackup locked all databases. If not, why the number of connections stayed on 512 even after killing the program?
Any help greatly appreciated.