[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

RE: [Linux-cluster] iscsi doubt



 

From: linux-cluster-bounces redhat com [mailto:linux-cluster-bounces redhat com] On Behalf Of ESGLinux
Sent: Monday, April 20, 2009 12:56 PM
To: linux clustering
Subject: Re: [Linux-cluster] iscsi doubt

I only want to share a directory in which one node writes at one and when it fails the other node has the diretory mounted with the data and can write to it.  
The ONLY safe way to do this with ext3 is to ensure the fs is mounted on one node at a time.
 
That means:
 
- The failover node cannot have the ext3 fs mounted until a failover is performed
- The failed node must not write to the fs after it fails (i.e. you fence the node)
- Prior to failing back to the primary node, the failover node must unmount the fs
 
I believe you can automate such a failover with rgmanager, but I have not done so myself.
 Before I have known about cluster my decission would been to mount the shares with NFS. Now I want to be more sofisticated and want to use cluster tools, so I thought to mount it with iSCSI instead of NFS, but always with the ext3 as the underlying filesystem.  
 iSCSI isn't a cluster tool, it's a block-level storage protocol.  Using iSCSI is neither necessary nor sufficient to implement a cluster.  You can use other software such as DRBD to get the failover you describe.
 
Jeff
 

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]