Hi, Could someone explain what happens if one of the OST fails? Is there any way to regenerate the data online? Or the only left way is to bring down the system and try to recover data manually? Thanks, Sumit -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20060919/fdee3125/attachment.html
Hi, You can configure OST nodes in failover mode, see https://mail.clusterfs.com/wikis/lustre/LustreFailover . For this it is important to have external attached storage, which can be accessed by a pair of OST servers. This storage has to be secured against hardware failure, for example using RAID5+Spare. Cheers, Jan -----Original Message----- From: lustre-discuss-bounces@clusterfs.com [mailto:lustre-discuss-bounces@clusterfs.com] On Behalf Of Sumit Narayan Sent: 19 September 2006 15:20 To: lustre-discuss@clusterfs.com Subject: [Lustre-discuss] What if OST fails? Hi, Could someone explain what happens if one of the OST fails? Is there any way to regenerate the data online? Or the only left way is to bring down the system and try to recover data manually? Thanks, Sumit -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20060919/5ca51d1f/attachment.html
There is a feature under development that we call LAID (Lustre RAID) - it introduces striping across OST targets and would allow rebuild of data. It''s a while out though before it is ready (2007). - Peter - _____ From: lustre-discuss-bounces@clusterfs.com [mailto:lustre-discuss-bounces@clusterfs.com] On Behalf Of Sumit Narayan Sent: Tuesday, September 19, 2006 8:20 AM To: lustre-discuss@clusterfs.com Subject: [Lustre-discuss] What if OST fails? Hi, Could someone explain what happens if one of the OST fails? Is there any way to regenerate the data online? Or the only left way is to bring down the system and try to recover data manually? Thanks, Sumit -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20060919/ffd70aff/attachment.html