Galera installation error...

Follow

Ron Peeters

• December 28, 2011 07:49

I'm unable to get the s9s installer going....

 

My Install log is attached

s9s_deploy.log
cc.log

0

• 

  Ron Peeters

  • December 28, 2011 07:53

  And I'm new to s9s installers with galera...

   

  Everything goes fine until the script tries to start mysql on the cluster control server...

  This is what the /var/lib/mysql/error.log reads:

  11228 08:34:41 mysqld_safe mysqld from pid file /var/lib/mysql/mysqld.pid ended

  111228 08:35:01 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql

  111228  8:35:01 [Note] Plugin 'FEDERATED' is disabled.

  111228  8:35:01 InnoDB: The InnoDB memory heap is disabled

  111228  8:35:01 InnoDB: Mutexes and rw_locks use GCC atomic builtins

  111228  8:35:01 InnoDB: Compressed tables use zlib 1.2.3

  111228  8:35:01 InnoDB: Using Linux native AIO

  111228  8:35:01 InnoDB: Initializing buffer pool, size = 5.6G

  InnoDB: mmap(6153453568 bytes) failed; errno 12

  111228  8:35:01 InnoDB: Completed initialization of buffer pool

  111228  8:35:01 InnoDB: Fatal error: cannot allocate memory for the buffer pool

  111228  8:35:01 [ERROR] Plugin 'InnoDB' init function returned error.

  111228  8:35:01 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.

  111228  8:35:01 [ERROR] Unknown/unsupported storage engine: InnoDB

  111228  8:35:01 [ERROR] Aborting

  111228  8:35:01 [Note] /usr/sbin/mysqld: Shutdown complete

  111228 08:35:01 mysqld_safe mysqld from pid file /var/lib/mysql/mysqld.pid ended

  0

  Comment actions Permalink
• 

  Johan

  • January 04, 2012 15:13

  Hi,

  Sorry for not having replied earilier but this one slipped through.

  It looks like you don't have enough memory to allocate to the innodb buffer pool.

  111228  8:35:01 InnoDB: Fatal error: cannot allocate memory for the buffer pool

  Check your servers memory, e.g, with the linux command
  free -m

  and generate a new package with the correct memory setting for the server.

  You can also do:

  vi s9s-galera-xyz/mysql/config/my.cnf

  and lower the innodb buffer pool size there, and then hit 'deploy.sh' again.

  Thank you,

  Johan

  0

  Comment actions Permalink
• 

  Bradley Watkins

  • January 09, 2012 21:06

  I think the issue, which I'm hitting now myself, is that there is an assumption that the cluster control system will have the same amount of memory as the Galera servers themselves.  I don't think this is an appropriate assumption.  In my case in particular, all of the Galera nodes are large-ish hardware nodes with 48GB of RAM while the control server is a VM with 4GB (which I think should be plenty).

  0

  Comment actions Permalink
• 

  Johan

  • January 10, 2012 06:08

  Hi, we have fixed that assumption now. Now you can set System Memory on the ClusterControl server.

  BR
  johan

  0

  Comment actions Permalink

Please sign in to leave a comment.