Re: big weird error - Adam Dixon - com.mysql.lists.cluster

14 messages in com.mysql.lists.clusterRe: big weird error

From	Sent On	Attachments
Matthew Story	21 Aug 2006 12:02
Adam Dixon	21 Aug 2006 16:22
Stewart Smith	21 Aug 2006 20:27
Matthew Story	23 Aug 2006 08:34
Matthew Story	23 Aug 2006 09:35
Matthew Story	23 Aug 2006 12:39
Matthew Story	23 Aug 2006 13:27
Adam Dixon	23 Aug 2006 17:20
Matthew Story	23 Aug 2006 17:49
Stewart Smith	23 Aug 2006 17:57
Adam Dixon	23 Aug 2006 17:57
Brendan Bouffler	23 Aug 2006 20:42
Brendan Bouffler	23 Aug 2006 20:54
Matthew Story	24 Aug 2006 08:06

Subject:	Re: big weird error
From:	Adam Dixon (adam...@gmail.com)
Date:	08/21/2006 04:22:30 PM
List:	com.mysql.lists.cluster

The ndbd processes will allocate the DataMemory and IndexMemory at start up, consuming it immediatly, so each datanode needs at least 2500MB of memory in it to start up without potential problems. Further to that, you can check CPU/Network IO on the machine to ensure that it is responsive, as if its under duress, it will/may fail checks that the ndbd process run.

Adam

On 8/22/06, Matthew Story <matt...@gmail.com> wrote:

Node 3: Forced node shutdown completed. Occured during startphase 0. Initiated by signal 0. Caused by error 6050: 'WatchDog terminate, internal error or massive overload on the machine running this node(Internal error, programming error or missing error m

I thought this might have to do with allocating too much memory, though at the time it began i was only allocating 2 GB data and .5 GB index, I have since stepped it down to a sad 150 MB data and 50 MB index, and I'm still getting the above error. This is kind of weird situation, because its not running on a subnet right now, its running accross multiple subnets. Any ideas? I'm running 5.0.20-max.