Please find the latest version of these slides at:

https://code.google.com/p/ganeti/wiki/LISA2013

• ... all start with gnt-
  • gnt-cluster
  • gnt-node
  • gnt-instance
  • ...
• ... all take a subcommand
  • gnt-cluster info
  • gnt-node list
  • gnt-instance list

gnt-[subsystem] [verb] --flags [noun]

• The man page
  • man gnt-instance
• List of subcommands
  • gnt-instance
  • gnt-instance --help
• Help on individual subcommands:
  • gnt-instance list --help

# gnt-instance list --help
Usage
=====
  gnt-instance list [<instance>...]

Lists the instances and their status. The available
fields can be shown using the "list-fields" command
(see the man page for details). The default field list
is (in order): name, hypervisor, os, pnode, status,
oper_ram.

Options
=======
--no-headers            Don't display column headers
--separator=SEPARATOR   Separator between output fields ...
...

Commands must run on the master. Any other node will give you a friendly message. Scripts can use "getmaster" to know the right place.

# gnt-node list
Failure: prerequisites not met for this operation:
This is not the master node, please connect to node
'gnta2.example.com' and rerun the command
# gnt-cluster getmaster
gnta2.example.com
# ssh gnta2.example.com
WARNING:
This machine is part of a ganeti cluster.
# gnt-node list
Node              DTotal DFree MTotal MNode MFree Pinst Sinst
gnta1.example.com   3.6T  3.1T  64.0G 1023M 15.0G     5     3
...etc...

gnt-cluster info
gnt-cluster modify [-B/H/N ...]
gnt-cluster verify
gnt-cluster master-failover
gnt-cluster command ...
gnt-cluster copyfile ...

# gnt-cluster verify
Submitted jobs 285450, 285451
Waiting for job 285450 ...
Sat Oct 27 19:14:08 2012 * Verifying cluster config
Sat Oct 27 19:14:08 2012 * Verifying cluster certificate files
Sat Oct 27 19:14:08 2012 * Verifying hypervisor parameters
Sat Oct 27 19:14:08 2012 * Verifying all nodes belong to an existing group
Waiting for job 285451 ...
Sat Oct 27 19:14:08 2012 * Verifying group 'default'
Sat Oct 27 19:14:08 2012 * Gathering data (3 nodes)
Sat Oct 27 19:14:10 2012 * Gathering disk information (3 nodes)
Sat Oct 27 19:14:11 2012 * Verifying configuration file consistency
Sat Oct 27 19:14:11 2012 * Verifying node status
Sat Oct 27 19:14:11 2012 * Verifying instance status
Sat Oct 27 19:14:11 2012 * Verifying orphan volumes
Sat Oct 27 19:14:11 2012 * Verifying N+1 Memory redundancy
Sat Oct 27 19:14:11 2012 * Other Notes
Sat Oct 27 19:14:11 2012   - NOTICE: 1 offline node(s) found.
Sat Oct 27 19:14:12 2012 * Hooks Results

gnt-node list
gnt-node info
gnt-node remove node4
gnt-node modify \
  [ --master-candidate yes|no ] \
  [ --drained yes|no ] \
  [ --offline yes|no ] node2
gnt-node evacuate/failover/migrate node3
gnt-node powercycle node1

# gnt-node list
Node              DTotal DFree MTotal MNode MFree Pinst Sinst
gnta1.example.com   3.6T  3.1T  64.0G 1023M 15.0G     5     3
gnta2.example.com   3.6T  3.1T  64.0G 1023M 22.9G     4     4
gnta3.example.com      *     *      *     *     *     0     0
gnta4.example.com   3.6T  3.1T  64.0G 1023M 21.0G     4     6

# gnt-node info gnta1
Node name: gnta1.example.com
  primary ip: 172.15.155.15
  secondary ip: 172.99.199.1
      ...etc...
  primary for instances:
    - ginny.example.com
      ...etc...
  secondary for instances:
    - webcsi.example.com
      ...etc...
  node parameters:
    - oob_program: default (None)
    - spindle_count: default (1)
      ...etc...

gnt-instance start/stop myinstance
gnt-instance modify ... myinstance
gnt-instance info myinstance
gnt-instance list
gnt-instance migrate myinstance
gnt-instance console myinstance

# gnt-instance list
Instance             Hypervisor OS            Primary_node      Status  Memory
rocker1.example.com  xen-pvm    debian-server gnta2.example.com running   512M
webcsi.example.com   xen-pvm    debian-server gnta3.example.com running   1.0G

# gnt-instance info rocker1
Instance name: rocker1.example.com
UUID: 3244567d-a08a-4663-8349-c68307fab664
Serial number: 2
Creation time: 2012-07-05 20:08:14
Modification time: 2012-07-09 15:33:03
State: configured to be up, actual state is up
  Nodes:
    - primary: gnta2.example.com
    - secondaries: gnta3.example.com
  Operating system: debian-server
  Allocated network port: None
  Hypervisor: xen-pvm
    - ...

• Ganeti operations generate jobs in the master
  • (with the exception of queries)
• Jobs execute concurrently
• You can
  • cancel non-started jobs,
  • inspect the queue status, and
  • inspect jobs
• Development in progress: more fine-grained job-queue management

gnt-job list
gnt-job info
gnt-job watch
gnt-job cancel

gnt-backup export -n node1 instance1
gnt-backup import -t plain \
  {-n node3 | -I hail } --src-node node1 \
  --src-dir /tmp/myexport instance1
gnt-backup list
gnt-backup remove

gnt-group add
gnt-group assign-nodes
gnt-group evacuate
gnt-group list
gnt-group modify
gnt-group remove
gnt-group rename
gnt-instance change-group

Customize gnt-* list output with -o:

# gnt-instance list -o name,snodes
Instance             Secondary_Nodes
rocker1.example.com  gnta3.example.com
webcsi.example.com   gnta2.example.com

--no-headers is useful in shell scripts:

# gnt-instance list -o name,snodes --no-headers
rocker1.example.com
webcsi.example.com

The list-fields subcommand lists all available fields.

gnt-group list-fields
gnt-instance list-fields
gnt-job list-fields
gnt-backup list-fields

Filter output of list subcommands using the -F option:

# gnt-instance list -F 'pnode == "gnta1"' --no-headers -o name
ringo.example.com
george.example.com
john.example.com
paul.example.com
luke.example.com

Filtering language is described in man ganeti

Examples:

'(be/vcpus == 3 or be/vcpus == 6) and pnode.group =~ m/^rack/'

'pinst_cnt != 0'

'not master_candidate'

Instances can be moved between clusters that share a common secret.

• Operation available via the CLI or RAPI
• CLI tool uses RAPI, and can be seen as an example
• Data is transferred directly between clusters

Setup common secret and RAPI authentication

ssh root@cluster1 --> root@cluster1:~#
gnt-cluster renew-crypto --new-cluster-domain-secret
cat > /var/lib/ganeti/rapi/users <<EOF
mover testpwd write
EOF

# copy /var/lib/ganeti/cluster-domain-secret to the second cluster

ssh root@cluster2 --> root@cluster2:~#
gnt-cluster renew-crypto --cluster-domain-secret=path_to_domain_secret
# rapi access can be the same or different. in production use hashed passwords.
cat > /var/lib/ganeti/rapi/users <<EOF
mover testpwd write
EOF

Can be run on a third party machine

PWDFILE=$(mktemp)
echo testpwd > $PWDFILE

# Note: --dst-* defaults to --src-* if not specified
/usr/lib/ganeti/tools/move-instance --verbose \
  --src-ca-file=rapi.pem --src-username=mover \
  --src-password-file=$PWDFILE \
  [--dest-instance-name=new_name --net=0:mac=generate] \
  --iallocator=hail cluster1 cluster2 instance.example.com

Bugs:

• Either--iallocator or nodes must be specified manually
• Move is slower than it ought to be
• Instances with file-based disk template not supported.