6 messages in com.mysql.lists.clusterRe: ndbd crash
FromSent OnAttachments
Adam Dixon18 Dec 2005 16:16 
Chad Martin18 Dec 2005 17:01 
Stewart Smith19 Dec 2005 15:47 
Adam Dixon20 Dec 2005 14:11 
Cor...@ SkyVantage20 Dec 2005 18:03 
Stewart Smith20 Dec 2005 19:03 
Subject:Re: ndbd crash
From:Stewart Smith (stew@mysql.com)
Date:12/19/2005 03:47:44 PM
List:com.mysql.lists.cluster

On Sun, 2005-12-18 at 19:01 -0600, Chad Martin wrote:

Adam Dixon wrote:

<ffffffffa003cfc8>{:jbd:journal_dirty_metadata+71} Trace:<ffffffffa00528c2>{:ext3:ext3_mark_iloc_dirty+740} Dec 18 03:18:09 data3 kernel: <ffffffffa0052a14>{:ext3:ext3_mark_inode_dirty+65} Dec 18 03:18:10 data3 kernel: <ffffffffa003c4c6>{:jbd:journal_start+223} <ffffffffa0054726>{:ext3:ext3_dirty_inode+99} Dec 18 03:18:10 data3 kernel: <ffffffff8018fef0>{__mark_inode_dirty+40} <ffffffff80189748>{inode_update_time+182} Dec 18 03:18:10 data3 kernel: <ffffffff80155b7b>{generic_file_aio_write_nolock+620} Dec 18 03:18:10 data3 kernel: <ffffffff80155c9c>{generic_file_aio_write+126} <ffffffffa004fed0>{:ext3:ext3_file_write+22}

All this stuff makes me think it's a corrupted filesystem. Have you tried running fsck on the filesystem to see if it's healthy?

I'd also report it to the ext3 people - they'll be able to help work out if it's in-memory corruption or not (or if it's a possible ext3/kernel bug).

Just think - thanks to cluster your DB stayed up even through an OS/Hardware error. Yay :)

-- Stewart Smith, Software Engineer MySQL AB, www.mysql.com Office: +14082136540 Ext: 6616 VoIP: 66@sip.us.mysql.com Mobile: +61 4 3 8844 332

Jumpstart your cluster: http://www.mysql.com/consulting/packaged/cluster.html