ceph/doc/install/manual-deployment.rst
John Wilkins de2bcd5da9 doc: Added a manual deployment doc.
Signed-off-by: John Wilkins <john.wilkins@inktank.com>
2013-11-20 19:27:47 -08:00

364 lines
12 KiB
ReStructuredText

===================
Manual Deployment
===================
All Ceph clusters require at least one monitor, and at least as many OSDs as
copies of an object stored on the cluster. Bootstrapping the initial monitor(s)
is the first step in deploying a Ceph Storage Cluster. Monitor deployment also
sets important criteria for the entire cluster, such as the number of replicas
for pools, the number of placement groups per OSD, the heart beat intervals,
whether authentication is required, etc. Most of these values are set by
default, so it's useful to know about them when setting up your cluster for
production.
Following the same configuration as `Installation (Quick)`_, we will set up a
cluster with ``node1`` as the monitor node, and ``node2`` and ``node3`` for
OSD nodes.
.. ditaa::
/------------------\ /----------------\
| Admin Node | | node1 |
| +-------->+ |
| | | cCCC |
\---------+--------/ \----------------/
|
| /----------------\
| | node2 |
+----------------->+ |
| | cCCC |
| \----------------/
|
| /----------------\
| | node3 |
+----------------->| |
| cCCC |
\----------------/
Monitor Bootstrapping
=====================
Bootstrapping a monitor (a Ceph Storage Cluster, in theory) requires
a number of things:
- **Unique Identifier:** The ``fsid`` is a unique identifier for the cluster,
and stands for File System ID from the days when the Ceph Storage Cluster was
principally for the Ceph Filesystem. Ceph now supports native interfaces,
block device, and object storage gateway interfaces too, so ``fsid`` is a bit
of a misnomer.
- **Cluster Name:** All Ceph clusters have a name, but the default name is
``ceph``. Setting the cluster name to something other than ``ceph`` is
especially useful when you are working with multiple clusters and you need to
clearly understand which cluster your are working with. For example, when you
run multiple clusters on the same hardware, or if you are running Ceph in
federated architectures, the cluster name defines which cluster you are
using (e.g., ``us-west``, ``us-east``).
- **Monitor Name:** Each monitor instance within a cluster has a unique name.
In common practice, the Ceph Monitor name is the host name (we recommend one
Ceph Monitor per host, and no commingling of Ceph OSD Daemons with
Ceph Monitors).
- **Cluster Map:** Bootstrapping the initial monitor(s) requires you to
generate a cluster map. The cluster map requires the ``fsid``, the cluster
name (or uses the default), and at least one host name and its IP address.
- **Monitor Keyring**: Monitors communicate with each other via a
secret key. You must generate a keyring with a monitor secret and provide
it when bootstrapping the initial monitor(s).
- **Administrator Keyring**: To use the ``ceph`` CLI tools, you must have
a ``client.admin`` user. So you must generate the admin user and keyring,
and you must also add the ``client.admin`` user to the monitor keyring.
The foregoing requirements do not imply the creation of a Ceph Configuration
file. However, as a best practice, we recommend creating a Ceph configuration
file and populating it with the ``fsid``, the ``mon initial members`` and the
``mon host`` settings.
You can get and set all of the monitor settings at runtime as well. However,
a Ceph Configuration file may contain only those settings that override the
default values. When you add settings to a Ceph configuration file, these
settings override the default settings. Maintaining those settings in a
Ceph configuration file makes it easier to maintain your cluster.
The procedure is as follows:
#. Log in to the initial monitor node(s)::
ssh {hostname}
ssh node1
#. Ensure you have a directory for the Ceph configuration file. By default,
Ceph uses ``/etc/ceph``. When you install ``ceph``, the directory will
get created automatically. ::
ls /etc/ceph
#. Create a Ceph configuration file (optional). By default, Ceph uses
``ceph.conf``, where ``ceph`` reflects the cluster name. ::
sudo vim /etc/ceph/ceph.conf
#. Generate a unique ID (i.e., ``fsid``) for your cluster. ::
uuidgen
#. Add the unique ID to your Ceph configuration file (optional). ::
fsid = {UUID}
fsid = a7f64266-0894-4f1e-a635-d0aeaca0e993
#. Add the initial monitor(s) to your Ceph configuration file (optional). ::
mon initial members = {hostname}[,{hostname}]
mon initial members = node1
#. Add the IP address(es) of the initial monitor(s) to your Ceph configuration
file and save the file (optional). ::
mon host = {ip-address}[,{ip-address}]
mon host = 192.168.0.1
#. Create a keyring for your cluster and generate a monitor secret key. ::
ceph-authtool --create-keyring /tmp/ceph.mon.keyring --gen-key -n mon. --cap mon 'allow *'
#. Generate an administrator keyring, generate a ``client.admin`` user and add
the user to the keyring. ::
ceph-authtool --create-keyring /etc/ceph/ceph.client.admin.keyring --gen-key -n client.admin --set-uid=0 --cap mon 'allow *' --cap osd 'allow *' --cap mds 'allow'
#. Add the ``client.admin`` key to the ``ceph.mon.keyring``. ::
ceph-authtool ceph.mon.keyring --import-keyring /etc/ceph/ceph.client.admin.keyring
#. Generate a monitor map using the hostname(s), host IP address(es) and the FSID.
Save it as ``/tmp/monmap``::
monmaptool --create --add {hostname} {ip-address} --fsid {uuid} /tmp/monmap
monmaptool --create --add node1 192.168.0.1 --fsid a7f64266-0894-4f1e-a635-d0aeaca0e993 /tmp/monmap
#. Create a default data directory (or directories) on the monitor host(s). ::
sudo mkdir /var/lib/ceph/mon/{cluster-name}-{hostname}
sudo mkdir /var/lib/ceph/mon/ceph-node1
#. Populate the monitor daemon(s) with the monitor map and keyring. ::
ceph-mon --mkfs -i {hostname} --monmap /tmp/monmap --keyring /tmp/ceph.mon.keyring
ceph-mon --mkfs -i node1 --monmap /tmp/monmap --keyring /tmp/ceph.mon.keyring
#. Consider settings for a Ceph configuration file. Common settings include
the following::
[global]
fsid = {cluster-id}
mon initial members = {hostname}[, {hostname}]
mon host = {ip-address}[, {ip-address}]
public network = {network}[, {network}]
cluster network = {network}[, {network}]
auth cluster required = cephx
auth service required = cephx
auth client required = cephx
osd journal size = {n}
filestore xattr use omap = true
osd pool default size = {n} # Write an object n times.
osd pool default min size = {n} # Allow writing n copy in a degraded state.
osd pool default pg num = {n}
osd pool default pgp num = {n}
osd crush chooseleaf type = {n}
In the foregoing example, the ``[global]`` section of the configuration might
look like this::
[global]
fsid = a7f64266-0894-4f1e-a635-d0aeaca0e993
mon initial members = node1
mon host = 192.168.0.1
public network = 192.168.0.0/24
auth cluster required = cephx
auth service required = cephx
auth client required = cephx
osd journal size = 1024
filestore xattr use omap = true
osd pool default size = 2
osd pool default min size = 1
osd pool default pg num = 333
osd pool default pgp num = 333
osd crush chooseleaf type = 1
#. Start the monitor(s).
For Debian/Ubuntu, use Upstart::
sudo start ceph-mon id=node1
For CentOS/RHEL, use sysvinit::
sudo /etc/init.d/ceph start mon.node1
#. Verify that Ceph created the default pools. ::
ceph osd lspools
You should see output like this::
0 data,1 metadata,2 rbd,
#. Verify that the monitor is running. ::
ceph -s
You should see output that looks something like this::
cluster a7f64266-0894-4f1e-a635-d0aeaca0e993
health HEALTH_ERR 192 pgs stuck inactive; 192 pgs stuck unclean; no osds
monmap e1: 1 mons at {node1=192.168.0.1:6789/0}, election epoch 1, quorum 0 node1
osdmap e1: 0 osds: 0 up, 0 in
pgmap v2: 192 pgs, 3 pools, 0 bytes data, 0 objects
0 kB used, 0 kB / 0 kB avail
192 creating
Adding OSDs
===========
Once you have your initial monitor(s) running, you should add OSDs. Your cluster
cannot reach an ``active + clean`` state until you have enough OSDs to handle the
number of copies of an object (e.g., ``osd pool default size = 2`` requires at
least two OSDs). After bootstrapping your monitor, your cluster has a default
CRUSH map; however, the CRUSH map doesn't have any Ceph OSD Daemons mapped to
a Ceph Node.
To create the first two OSDs, perform this procedure on ``node2`` and ``node3``:
#. Generate a UUID for the OSD. ::
uuidgen
#. Create the OSD. If no UUID is given, it will be set automatically when the
OSD starts up. The following command will output the OSD number, which you
will need for subsequent steps. ::
ceph osd create [{uuid}]
#. Create the default directory on your new OSD. ::
ssh {new-osd-host}
sudo mkdir /var/lib/ceph/osd/ceph-{osd-number}
#. If the OSD is for a drive other than the OS drive, prepare it
for use with Ceph, and mount it to the directory you just created::
ssh {new-osd-host}
sudo mkfs -t {fstype} /dev/{drive}
sudo mount -o user_xattr /dev/{hdd} /var/lib/ceph/osd/ceph-{osd-number}
#. Initialize the OSD data directory. ::
ssh {new-osd-host}
sudo ceph-osd -i {osd-num} --mkfs --mkkey
The directory must be empty before you can run ``ceph-osd``.
#. Register the OSD authentication key. The value of ``ceph`` for
``ceph-{osd-num}`` in the path is the ``$cluster-$id``. If your
cluster name differs from ``ceph``, use your cluster name instead.::
sudo ceph auth add osd.{osd-num} osd 'allow *' mon 'allow rwx' -i /var/lib/ceph/osd/ceph-{osd-num}/keyring
#. Add your Ceph Node to the CRUSH map. ::
ceph osd crush add-bucket {hostname} host
ceph osd crush add-bucket node1 host
#. Place the Ceph Node under the root ``default``. ::
ceph osd crush move node1 root=default
#. Add the OSD to the CRUSH map so that it can begin receiving data. You may
also decompile the CRUSH map, add the OSD to the device list, add the host as a
bucket (if it's not already in the CRUSH map), add the device as an item in the
host, assign it a weight, recompile it and set it. ::
ceph osd crush add {id-or-name} {weight} [{bucket-type}={bucket-name} ...]
ceph osd crush add osd.0 1.0 host=node1
#. After you add an OSD to Ceph, the OSD is in your configuration. However,
it is not yet running. The OSD is ``down`` and ``in``. You must start
your new OSD before it can begin receiving data.
For Debian/Ubuntu, use Upstart::
sudo start ceph-osd id={osd-num}
sudo start ceph-osd id=0
sudo start ceph-osd id=1
For CentOS/RHEL, use sysvinit::
sudo /etc/init.d/ceph start osd.{osd-num}
sudo /etc/init.d/ceph start osd.0
sudo /etc/init.d/ceph start osd.1
Once you start your OSD, it is ``up`` and ``in``.
Summary
=======
Once you have your monitor and two OSDs up and running, you can watch the
placement groups peer by executing the following::
ceph -w
To view the tree, execute the following::
ceph osd tree
You should see output that looks something like this::
# id weight type name up/down reweight
-1 2 root default
-2 2 host node1
0 1 osd.0 up 1
-3 1 host node2
1 1 osd.1 up 1
To add (or remove) additional monitors, see `Add/Remove Monitors`_.
To add (or remove) additional Ceph OSD Daemons, see `Add/Remove OSDs`_.
.. _Installation (Quick): ../../start
.. _Add/Remove Monitors: ../../rados/operations/add-or-rm-mons
.. _Add/Remove OSDs: ../../rados/operations/add-or-rm-osds