Re: Memory exhausted on one node
Hi Jonathan,
Thank for your help.
1) I am using MySQL-Max distrib 4.1.11, for pc-linux-gnu (i686).
2) The system of ndb_mgmd is the same as ndbd nodes: Fedora core3(2.6.11-1.14_FC3).
3) When I stoped the exhausted mysqld(ndbd too) by mysqladmin shutdown. It took a very long time(more than 20 minutes), then one table has some problem(Maybe index problem, because if I select it with 'where' it give no response, but without 'where' it works ok), I found the error messages in log file:
050602 14:32:00 [ERROR] Got error 4012 when reading table './radius/radcheck'
050602 14:32:39 [ERROR] Got error 4010 when reading table './radius/radcheck'
050602 14:32:39 [ERROR] Got error 4010 when reading table './radius/radcheck'
050602 14:32:39 [ERROR] /usr/local/mysql/bin/mysqld: Can't open file: '#sql-35b4_578.ndb' (errno: 1)
After I shutdown entire cluster, this table works good now.
4) I established a test platform, it seems no problem. maybe the workload of test platform is not heavy enough?
5) All three nodes has the same OS version and mysql version.
The only diffrence is: tow ndb+mysql nodes are using Dell server. but Mgr node is using a less powerful pc server.
Another thing I should mention is: We put a load balancer befor to mysqld nodes, and it works smoothly now(it really distribute load in two equal part). Is there any problem?
The work load of current system is not heavy now, I am worrying about the future!
Thanks a lot!