Procedure
1. Log on to the SMW as root.
2. Stop the name nodes.
#
urika-stop
For more information, refer to the
urika-stop
man page.
3. Remove the tenant VMs
#
ux-tenant-remove --vm-only tenant-name
For more information, refer to the
ux-tenant-remove
man page.
4. Back up data/files.
While the HDFS data on Urika-GX is spread across data nodes and is stored redundantly to prevent loss, the
name nodes for both tenant and non-tenant data store the metadata on NID 0. This means that if NID 0 loses
its data, the user data may persist, but the ability to identify and retrieve it will be lost. To prevent this when
swapping, redeploying or wiping NID 0, it is necessary to bring all name nodes to a quiescent (
shutdown
)
state, and then copy the data found in the directory tree
/mnt/hdd-2/hdfs
to a safe place. Shutting down
the name nodes is achieved by running the
urika-stop
command. The operating environment for any given
tenant VM resides in a disk image file found on the host node (the node named in
the
/etc/sysconfig/uxtenant/hosts/
host-name
file named in the tenant configuration)
at
/qemu/
tenant-name
.img
.
a. Copy the files from
/mnt/hdd-2/hdfs
on NID 0 to a safe place.
b. Copy the tenant disk image(s) from
/qemu/
tenant-name
.img
on the affected host node(s) to a safe
place.
5. Perform the required operations on the node(s).
6. Bring the nodes back up, making sure they are completely deployed and integrated. Contact Cray Support for
guidance on node redeployment.
7. Put back the name node data on NID 0 at
/mnt/hdd-2/hdfs
8. Put back the tenant VM disk image files on the affected host nodes.
9. Restore the tenant VMs.
#
ux-tenant-create --use-image-file path tenant-name
In the preceding command,
path
is the path to the disk image file on the host node where the disk image is
put back. For more information, refer to the
ux-tenant-create
man page.
10. Restart the name nodes
#
ux-tenant-start
For more information, refer to the
ux-tenant-start
man page.
Troubleshooting
S3016
255
