NFS (Network File System) provides a convenient way to attach your Third Light Chorus server to a remote file store, often referred to as NAS (Network Attached Storage). NFS is not offered as a partitioning option during the installation stages, and should generally only be used when local disk storage is not an option.
To use NFS, you must first install your server to the local operating system drives, and then replace /space with an NFS mount. Once the NFS partition is mounted it will behave exactly as a normal partition, but if connectivity to the NFS server is lost your server will “lock” until the NFS server connection is restored.
When migrating content it is essential to preserve permissions and the existence of hidden files. 'root_squash' options must be disabled on the NFS server.
To use NFS, log in as root and follow these steps:
Install the NFS client
apt-get install nfs-client
Move /space to /space_old
mv /space /space_old
Note: if /space is a mounted partition, not a directory, then you should instead umount /space and remount the partition to /space_old. The easiest way to do this is by editing /etc/fstab and then running mount /space_old.
Make a new /space partition
mkdir /space chown ims:ims /space
Prepare to mount the NFS partition
Edit /etc/fstab to mount the /space partition. An example NFS mount table entry is as follows:[nfs-server]:/[path-to]/space /space nfs rw,noatime,rsize=32768,wsize=32768 0 0
[nfs-server] should be replaced with the IP address of your NFS host, eg. 192.168.0.10. The value of [path-to] is the remote path to your storage on the NFS server. For example:
192.168.0.10:/thirdlight/space /space nfs rw,noatime,rsize=32768,wsize=32768 0 0
Mount the NFS partition
mount /space
Restore the content of /space
mv /space_old/* /space_old/.imsdatastore /space
Reboot:
shutdown -r now
ISCSI SupportiSCSI provides emulation of the SCSI protocol to mount a remote disk via a fast network connection. This is generally the preferred option when using a SAN (Storage Area Network).
This article is intended for expert users.
iSCSI on Physical ServersiSCSI configuration is completed before the Third Light Chorus installation ISO is booted.
The network interfaces on most modern servers support operation as iSCSI initiators, configured via the system BIOS. Hardware iSCSI HBAs are configured in the same way, but typically under their own initialisation during the boot process.
To use iSCSI with your Chorus server, you should configure the host server's iSCSI initiator to connect to the iSCSI target, and thus expose the drive as a LUN to the OS before the installation stages. Then, in the partitioning stages of installing Chorus, your iSCSI target will appear as a device (typically as the second SCSI adapter, ie. /dev/sdb or possibly /dev/sdc). Format this as XFS and mount it as /space.
There is no additional configuration required on your Third Light server; iSCSI allows you to format the partition as if it were a local disk.
iSCSI on Virtual Machines
VMware vSphere
Under vSphere, RDM (raw device mappings) can be used to attach iSCSI (or FC, etc.) LUNs to virtual machines, up to a maximum size of 2TB.
Using this configuration, the iSCSI protocol is handled by the host, and the virtual machine sees an attached SCSI device, equivalent to a VMDK-based virtual disk. This method offers the simplest configuration, as the iSCSI LUN is partitioned and mounted during the installation process (see the section on iSCSI with physical servers above). This method also makes it possible to use hardware acceleration if the host has a hardware iSCSI HBA.
If you prefer not to use RDM, or you intend to allocate LUNs in excess of 2TB, see the software configuration process in the "Other Hypervisors" section below.
Other Hypervisors
These instructions can be used in any virtual or hardware environment, but the methods above are usually preferred if available.
First, install the software iSCSI initiator tools:
apt-get update apt-get install open-iscsi echo "node.startup = automatic" >> /etc/iscsi/iscsid.conf /etc/init.d/open-iscsi restart
Next, find the initiator name:
cat /etc/iscsi/initiatorname.iscsi
You can either use this initiator name to set up the ACLs on your iSCSI target, or edit this file to change the initiator name to one of your choosing.
To list the available targets, run the following (in this example, the target server is at
iscsi-target.example.comiscsiadm -m discovery -t st -p iscsi-target.example.com
This outputs a list of portal/target IQN for visible targets. If the desired target is not visible, re-check the ACL configuration and then re-run this command.
For this example, we will assume that the target IQN is
iqn.2013-08.com.example:iscsi.ims1If you use CHAP authentication, you can configure it as follows:
iscsiadm -m node --targetname "iqn.2013-08.com.example:iscsi.ims1" --portal "iscsi-target.example.com:3260" --op=update --name node.session.auth.authmethod --value=CHAP iscsiadm -m node --targetname "iqn.2013-08.com.example:iscsi.ims1" --portal "iscsi-target.example.com:3260" --op=update --name node.session.auth.username --value=username iscsiadm -m node --targetname "iqn.2013-08.com.example:iscsi.ims1" --portal "iscsi-target.example.com:3260" --op=update --name node.session.auth.password --value=password
To attach the LUN(s), run:
iscsiadm -m node --targetname "iqn.2013-08.com.example:iscsi.ims1" --portal "iscsi-target.example.com:3260" --login /etc/init.d/open-iscsi restart
The LUNs associated with the target will now be available as
/dev/sdX-fdisk -lwill help to identify them.You can now format and mount the new storage as required.
When adding to /etc/fstab, specify the
_netdevoption, to ensure that it is mounted after the iSCSI startup completesIf you attach additional LUNs to the target, you can either reboot the server or run
/etc/init.d/open-iscsi restartto make them available. To add additional targets, repeat the steps above from the discovery section.Using XFS is preferred as XFS is a growable filesystem.
