Commit Graph

61 Commits (d24decd8713b89f1a92d927d875cd0cd61d67b46)

Author SHA1 Message Date
Sébastien Han e9360456f0 Add more option to maximise sequential IO
Using the largeio,inode64,swalloc options usually improves sequential
workloads.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-15 10:54:08 +01:00
Sébastien Han d8ab8c0bd3 Add yum priority plugin
Without this plugin if a Ceph version is present in a repo (let's say
epel) it will install the epel version and not the ICE version.
We install yum-plugin-priorities.noarch to honor the 'priority=1' flag.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-13 15:46:47 +01:00
Sébastien Han 25cf4d8650 Ability to disable transparent hugepage
In storage world it's often recommended to disable transparent hugepages
as they will tend to lower performance.
Note that this change won't survive reboot. There are several ways to
disable this permanently such as:

* rc.local
* grub boot line

It's a bit tricky to do this in Ansible since it really depends on the
OS you're running on.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-13 11:11:11 +01:00
Sébastien Han 2cb8941dcc Restructure OS tuning and add new value
We now use a single task and loop over a set of params, it's easier to
add now params.

Setting zone_reclaim_mode to 0.
See the discussion on the Ceph ML:
http://lists.ceph.com/pipermail/ceph-users-ceph.com/2013-December/036211.html

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-13 10:34:46 +01:00
Sébastien Han 81b44fcfe2 Make OS tuning var configurable and file re-ordering
Depending on the OS you are runnning on you should be able to configure
these values.
Re-ordering file for clarity as well.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-07 11:44:40 +01:00
Sébastien Han 5e6708394a Fix ICE installation
Wrong ordering and variable.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-07 11:25:34 +01:00
Sébastien Han 5c66801dc7 Bump up maximum number of open files
Large cluster will easily reach the default value so we increase it.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-06 15:26:44 +01:00
Sébastien Han 79f6415e9d Bump up max open file for big cluster
Big cluster will easily reach the default limit so we need to increase
it and make it configurable.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-06 14:23:23 +01:00
Sébastien Han 08330b1fab Move role variables to defaults
Following best practices from
http://docs.ansible.com/playbooks_variables.html#variable-precedence-where-should-i-put-a-variable

Closes: #161

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2015-01-06 00:13:10 +01:00
Boris Manojlovic 33b8603d6e add quotes so ansible yaml parser do not barf on it... 2014-12-24 17:07:50 +01:00
Sébastien Han a796a962ef Add ICE Kernel module support for Red Hat systems
Install Kernel modules from ICE for Ceph RBD.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-12-18 12:25:17 +01:00
Sébastien Han 24e2ad8e50 Add multi-OS support for ICE
ICE supports Ubuntu, CentOS and RHEL.
Adding those.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-12-18 11:10:14 +01:00
Sébastien Han 81fe11fae5 Disable MDS and RGW from default
MDS and RGW are not deployed often (RGW more), so we disable them from
the default deployment to only get MONs and OSDs.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-12-16 14:52:49 +01:00
Sébastien Han db91b20308 Add support for Intank Ceph Enterprise
With the appropriate subscription details you will be able to use the
Inktank Ceph Enterprise version of Ceph running on RHEL7.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-12-16 14:47:35 +01:00
Leseb 93d86cd938 Merge pull request #151 from matthewrees/100-continue
Support for 100-continue toggling for RGW
2014-11-26 20:12:17 +01:00
Sébastien Han 621df22a4a Fix auto uuid generation
Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-14 15:21:30 +01:00
Matthew Rees accbfeff31 Support for 100-continue toggling for RGW
Tested for Precise and Trusty.
2014-11-14 16:05:56 +02:00
Leseb 3a485bf22a Revert "Fix autogen UUID and add auto gen monitor key" 2014-11-14 15:03:50 +01:00
Sébastien Han a7bba85027 Fix autogen UUID and add auto gen monitor key
This fixes the UUID auto generation and introduces the monitor key auto
generation.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-13 15:38:10 +01:00
Sébastien Han a107063543 Auto generate an FSID
It has becomes really anoying to manually generate an fsid prior to the
inital bootstrap. This commit introduces a method that auto-generates an
fsid. If for whatever reasons you want to force your own fsid you can
simply edit these 3 files and override the fsid variable:

- roles/ceph-common/vars/main.yml
- roles/ceph-mon/vars/main.yml
- roles/ceph-osd/vars/main.yml

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-13 14:08:27 +01:00
Leseb cc4395ae17 Merge pull request #145 from matthewrees/separate-ceph-keys
Standardising key addition
2014-11-13 13:48:02 +01:00
Matthew Rees e2d6e66440 Revert URL-based key addition for Debian 2014-11-13 14:11:51 +02:00
Sébastien Han ef7891fe84 Fix wrong variable for cephx signature
Add a comment for KRBD, enabling cephx signatures will prevent you to
map RBD devices.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-13 10:59:42 +01:00
Matthew Rees b6fcd37bc0 Standardising key addition
Debian based repo’s now fetch keys in a similar manner to Redhat based
ones, and both stable and dev keys have entries in vars.
2014-11-11 17:10:32 +02:00
Leseb cbb878beba Merge pull request #144 from leseb/ceph-client
Add more detail for the client socket
2014-11-11 12:02:36 +01:00
Sébastien Han 1d8871e615 Add more detail for the client socket
Adding: cluster name, type, id and cctid.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-11 12:01:31 +01:00
Leseb c8f2f6610e Merge pull request #141 from matthewrees/monitor-interface
monitor_interface template change
2014-11-11 10:55:41 +01:00
Sébastien Han a0b659e66b Fix wrong vars during RH install
Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-11 10:50:03 +01:00
Matthew Rees ec16fe31f3 monitor_interface template change
This will fix the monitor_interface issue for Vagrant provisioning
while being preferred to using group_var entries for physical
deployments.
2014-11-10 16:28:13 +02:00
Leseb 547f16b7aa Merge pull request #140 from leseb/rbd-cache
Add a client section in the ceph configuration file
2014-11-10 14:52:11 +01:00
Sébastien Han 29ea4e6f47 Add a client section in the ceph configuration file
In this section we enabled the rbd cache and a socket per client.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-10 14:50:19 +01:00
Sébastien Han 51070cc4fb Use Giant version
Use the last stable version: Giant

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-10 14:34:45 +01:00
Sébastien Han 5e52d89e94 Increase kernel.pid_max to a higher value
While running big boxes with 72 disks it's easy to get out of PID for
all the threads needed by Ceph. Increasing the default value removes
this limitation.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-11-10 14:13:31 +01:00
Sébastien Han 74ede6e166 Merge branch 'master' of https://github.com/jjoos/ceph-ansible into jjoos-master
Signed-off-by: Sébastien Han <sebastien.han@enovance.com>

Conflicts:
	roles/ceph-common/defaults/main.yml
	roles/ceph-common/tasks/Debian.yml
	roles/ceph-osd/tasks/journal_collocation.yml
	roles/ceph-osd/tasks/osd_directory.yml
	roles/ceph-osd/tasks/raw_journal.yml
	roles/ceph-osd/tasks/raw_multi_journal.yml
2014-11-06 15:18:56 +01:00
Sébastien Han 3a083ce1b2 Remove ansible managed header from ceph.conf
Fixes: #125

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-10-10 15:41:31 +02:00
Leseb c57f875a75 Merge pull request #110 from leseb/fix-updatedb
Fix updatedb disable
2014-10-10 15:36:20 +02:00
Leseb e0b9aac9b2 Merge pull request #127 from leseb/remove-ansible-managed-from-ceph-conf
Remove ansible managed header from ceph.conf
2014-10-10 15:35:09 +02:00
Sébastien Han f34e60813b Remove ansible managed header from ceph.conf
In ceph-common you load {{ ansible_managed }} at the top of the main
config file - this will trigger handlers on that file whenever an
Ansible run is made.

I'd suggest replacing it with a vanilla text comment 'managed by
Ansible' to warn
admins but avoid unnecessary cluster bounces.

fixes: #125

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-10-10 15:34:11 +02:00
Gauvain Pocentek 3386578954 Fix the mon.* generation in ceph.conf
The ceph.conf.j2 template currently always uses the current host facts
to get the IP address of each host in the mon loop. This is not the
expected behavior. This patch uses the correct facts to get the IP.
2014-10-10 15:27:15 +02:00
Sébastien Han 07216bf883 Add more recovery options
Recovery and/or re-balancing decrease performance, adding more options
might help tweaking this behavior.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-09-28 17:31:50 +02:00
jjoos 5e8141f8cf Sweep of all sorts of things. 2014-09-04 21:14:11 +02:00
Sébastien Han 6413b9c251 Improve RedHat support
Fix the rpm key and the repos integration.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-09-04 12:18:17 +02:00
Sébastien Han 83be67cd54 Use a very unlikely subnet address for Vagrant
Since 192.168.0.0/24 is very commong and might overlap with some
existing networks on your laptop, using another subnet like '42' is less
bound to happen.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-09-03 00:13:07 +02:00
Sébastien Han d0403b820e Fix OSD update during startup
Proviously we used osd_crush_update_on_start: true, this was interpreted
by Ansible as a boolean and appeared as 'True' inside the Ceph configuration
file. However the Ceph's init script is looking for 'true'.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-09-03 00:04:23 +02:00
Sébastien Han 4196de25d4 Add developement packages support
This commits introduces the support of the developpement branches of
Ceph. You can now install Ceph from master.
The behavior is done through 2 new options:

* ceph_stable: true will use the stable branch
* ceph_dev: true will use the dev branch

For the dev packages don't forget to set the branch that you want to
use.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-09-02 02:37:18 +02:00
Sébastien Han 7e591a4cce Fix some missing option declaration
Missing decleration and cosmetic.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-08-30 02:58:27 +02:00
Sébastien Han b81392d1c9 Fix restart on config change handler
Prior to this patch, the first match was winning and the playbook wasn't
doing any difference both "restart ceph", adding a distro filtrer fixes
this.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-08-30 02:34:11 +02:00
Sébastien Han ec6d4ddaba Fix updatedb disable
In some system the command doens't exist and fail to run the playbook.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-08-30 02:29:07 +02:00
Sébastien Han e2e774d516 Add more common options
Extend settings support.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-08-30 02:26:40 +02:00
Sébastien Han b3889c4442 Disable updatedb OSD directory parsing
It has been reported a couple of months ago by Dan van der Ster from
CERN that updatedb was consumming 100% of CPU while parsing system's
directories. Indeed the process was parsing the OSD PG directories that
might contains billions of objects.

Signed-off-by: Sébastien Han <sebastien.han@enovance.com>
2014-08-21 14:21:05 +02:00