| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
| |
Modify the erlang_cookie for root's home directory in the initrd script
to enable the root user to run rabbitmq configuration options.
Signed-off-by: Amy Fong <amy.fong@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Memory leak happens when the dynamic list times grows without anything to
reset it.
In ceilometer and cinder configuration files, the new option is created:
[nova_client]
max_timing_buffer=<value>
In all clients found that uses extends the HTTPClient and uses the times
list, we limit the size of the list by popping off the oldest item in the
list to maintain a maximum size. A default size of 200 is chosen,
configurable by the above configuration option.
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
| |
Explicitly enable --log-dir to enable logging where available
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
argparse is part of python 2.7+, so there's no need to install the .egg
version of argparse.
Nova was pulling in this requirement, which doesn't cause a problem until
another python app includes both argparse and the built in one.
That results in the following:
UserWarning: Module argparse was already imported from /usr/lib/python2.7/argparse.pyc, but /usr/lib/python2.7/site-packages is being added to sys.path
So we drop the argparse dependency, everything still works and we no longer
have the warning.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Updating to the R16B erlang version from: https://github.com/joaohf/meta-erlang
This is a cherry pick of the latest to avoid adding a new layer dependency.
Eventually, this will be dropped and the layer used directly.
We also update rabbitmq-server to depend on erlang-modules to get the proper
support on target.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
| |
To ensure that the free space required for proper operation is not
clobber by other layers manipulating the extra space variable, we make
our additions additive versus absolute.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
| |
The existing version of rabbitmq is showing its age, updating it to 3.2.4
so we can take advantage of better debug and new functionality.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
| |
Usage info for initrd script says status twice, cleaning up.
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
| |
Add the extras and debug package groups by default. These can later be
convered to image or distro features, but for now, we turn them on
universally.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
| |
Introducing the "extras" package group that hold non-essential, but packages
that increase the usability of the cloud userspace. These increase the size
of the image, but we value usability over small size at this point.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
| |
Add a package group which contains tools and utilities for debugging a
cloud system.
This currently only contains the core-tools packagegroup, but cloud
specific utilities will be added over time.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
| |
xterm's configure uses freetype, but there is no dependency on it in
the recipe. This is a race condition that leads to host contamination
and QA failures if the host's freetype-config is used.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Starting : FAILED - check /var/log/rabbitmq/startup_{log, _err}
rabbitmq-server status also shows an error message:
Status of node rabbit@controller ...
Error: unable to connect to node rabbit@controller: nodedown
<snip>
There's a problem with erlang cookie synchronization.
rabbitmq's home directory is created and is used as the location
for the cookie.
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
| |
tgtd implements status and reload, need to add this to the usage info
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
| |
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
| |
Create a README.spice and document and special consideration, usage....
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
| |
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
| |
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
| |
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We modify the controller so that both vnc and spice html5 proxies are running.
Since in the current recipe, both controller and compute share the same nova.conf
file, in the compute, both vnc and spice html5 are set to enabled and in this
configuration, vnc is chosen (in source: vnc is chosen because it's better
tested).
To change to spice, only change required is on the compute side.
The controller (since the default console_type is auto) will try to talk
to both and run whichever is available.
Signed-off-by: Amy Fong <amy.fong@windriver.com>
|
|
|
|
| |
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
nova-novncproxy and nova-spicehtml5proxy both do not check to see if the process
specified in PIDFILE exists before firing off a new one. Both initrd files can
actually use nova.init (used for the many other nova-* daemons)
nova-consoleauth also has this issue, we add a check in start to see if the process
in PIDFILE exists and exits if it does.
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Package spice-html5
Modify python-nova to add infrastructure for nova-spicehtml5proxy
support.
Note: spice is the default
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
| |
The "default_store" option in /etc/glance/glance-api.conf
controls how glance store images. Make this option configurable
through variable GLANCE_DEFAULT_STORE.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
|
|
|
|
|
|
|
|
| |
Add a convenient script for creating Cinder
volume types which are currently supported.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Enable nova-compute to:
* use cinder volume stored in a ceph pool as a block device
* store glance image into a ceph pool.
Also port 2 patches from https://github.com/openstack/nova
branch master into Havana branch.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
|
|
|
|
|
|
|
|
| |
Enable cinder-backup to use ceph as backend. It mainly
changes cinder.conf and creates cinder-backup package.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
|
|
|
|
|
|
|
| |
Add ceph rbd backend into multi-backend cinder-volume support.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
|
|
|
|
| |
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
| |
Commit 91c372f287732cdedbd7c1204c6ba5f34e5b93f6 [core-image-basic /
packagegroup-core-basic: rename to *-full-cmdline] in yocto renamed
these tasks so we need to reflect that change here.
Signed-off-by: Mark Asselstine <mark.asselstine@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
| |
Commit 0460bd62663d6db4c9b6aed5edd73fb0d052ffb2 [lvm2: allow compiles
to use latest readline] forced us to use an old version of
readline. Here we backport an upstream lvm2 commit which allows us to
use new readline and drop the PREFERRED_VERSION_ for readline.
Signed-off-by: Mark Asselstine <mark.asselstine@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
| |
In order for our bbappends to be effective we need version match with
what is in meta-virtualization.
Signed-off-by: Mark Asselstine <mark.asselstine@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
| |
The sql-alchemy project page has remove the 0.7.9 tarball. So we switch
to the pypi version.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
| |
Add status/reload to sysvinit scripts
Modify tgtd to make start/stop work better (borrowed from Debian's
implementation)
Signed-off-by: Amy Fong <amy.fong@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
| |
To support the cinder iscsi backend, we need iscsi-* recipes. The HIPOS
layer contains a reference implementation, but has distro bindings that
we must avoid.
So for now, we make a copy and reference the other layer.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
| |
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
| |
readline 6.x triggers build errors with LVM2, so we need to set
our preferred version to 5.x.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
| |
By default lvm-iscsi uses loopdev as lvm physical volume.
This patch makes it a bit easier to allow other layer to
be able configure lvm physical volume.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently all the openstack components have default start level
of 20. There are other services such as glusterfs, rabbbitmq,
database... are also starting at the same start level. On some
platform, this can cause racing condition between services which
in turn causes some of openstack components not started.
By adjusting the openstack components start level to higher will
ensure that system services start in the determistic way.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch adds glusterfs backend into multi-backend cinder
support (total of lvm-iscsi, nfs, and glusterfs). This including
changing cinder.conf, creating missing glusterfs_shares config file,
and adding glusterfs-client into compute/controller node and
gluster-server into controller node.
Also meta-filesystems is included to pickup fuse filesytem which
is needed by glusterfs.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
| |
This patch enables multi-backend support (lvm-iscsi and nfs) on
cinder-volume. This including changing cinder.conf, creating
missing nfs_shares config file, and adding mount.nfs into compute
node image.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
| |
Currently compute-note does not have initiator iscsi working
which is required to mount cinder block device over iscsi.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
| |
It's needed to implicitly activate the lvm
logical group at startup. Otherwise, cinder lvm
logical volume will not be created.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
| |
Every time cinder-volume is restarted with
"/etc/init.d/cinder-volume restart" a new loop dev
is created for lvm volume backing file instead of
reuse the existing loop dev
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
| |
On controller node, cinder requires qemu-img util to create
a volume based on glance image (for qcow2 disk-format).
So this patch adds qemu package into the controller image.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
In order for Libvirt compute node to be able to mount Cinder LVM
backend driver volume (hosted by Cinder Volume service) over iscsi,
the initiator iscsi kernel-module is needed.
This patch simplies to add kernel cfg for turning on iscsi-tcp kernel
config.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
| |
Allow backing file size for LVM backend driver (for cinder volume
serivce) to be configurable.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
| |
Allow the final controller node image rootfs extra space
to be configurable.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This patch creates bbappend for this bb file
git://github.com/DFE/HIPOS/meta-hidav/recipes-connectivity/open-iscsi/open-iscsi-user_2.0-871.bb
To build iscsi initiator (open-isisc) usable by openstack Cinder.
Signed-off-by: Vu Tran <vu.tran@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|
|
|
|
|
|
|
| |
Adding new recipe for glusterfs package.
Signed-off-by: Donn Seeley <donn.seeley@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
|