Re: monitoring of NDB parameters - Simon Garner - com.mysql.lists.cluster

10 messages in com.mysql.lists.clusterRe: monitoring of NDB parameters

From	Sent On	Attachments
Walter Bertot	15 Dec 2005 19:20
Walter Bertot	16 Dec 2005 10:10
Mikael Ronström	16 Dec 2005 12:48
Simon Garner	18 Dec 2005 14:04
Alex Davies	18 Dec 2005 14:44
Simon Garner	18 Dec 2005 15:19
Adam Dixon	18 Dec 2005 15:42
Jonathan Miller	19 Dec 2005 04:41
Stewart Smith	19 Dec 2005 15:31
Stewart Smith	19 Dec 2005 15:38

Subject:	Re: monitoring of NDB parameters
From:	Simon Garner (sgar...@expio.co.nz)
Date:	12/18/2005 02:04:51 PM
List:	com.mysql.lists.cluster

As a semi-related question: Is it necessary to restart all the NDB nodes (in addition to the MGM node) after changing any of the parameters in the MGM node's config file?

I've been assuming it is, but it would be nice if it weren't. Restarting all the NDBs becomes quite a chore with a large database.

-Simon

On 17/12/2005 9:49 a.m., Mikael Ronström wrote:

Hi, Unfortunately there is little support right now for receiving warnings. They exist for running out of IndexMemory and DataMemory but not for most of the others. It's a good suggestion however and will be taken into consideration. However this particular parameter is exhausted when big transactions are executed and so the warning system would most likely be too slow to detect it. A development that currently is ongoing and will arrive in some future release is to use a common pool for all those specific memory segments and that will be configurable and that one it makes good sense to issue warnings although large transactions will still be an issue to handle.

Rgrds Mikael

2005-12-16 kl. 19.10 skrev Walter Bertot:

this morning i received the following error in our logs

"ERROR.*Out of index operations in transaction coordinator (increase MaxNoOfConcurrentIndexOperations)"

While increasing this value did resolve our issue it would have been better for us to have the ability to receive a warning allowing us to be proactive rather than reactive to this type of error.

I can not find any documentation which show how we can see the present value of MaxNoOfConcurrentIndexOperations or for that matter any other parameter we configure for NDB.

Is there any way to see these values? Perhaps via the manager?

Mikael Ronstrom, Senior Software Architect MySQL AB, www.mysql.com

Jumpstart your cluster: http://www.mysql.com/consulting/packaged/cluster.html My blog: http://mikaelronstrom.blogspot.com