The copy image-to-volume errors weren't being handled
properly and the result was the lvcreate being retried
even though the lvcreate itself succeeded.
The result of this was misleading errors stating that
the volume couldn't be created because it already existed
(which it did, becuase the create itself was succesful).
Fixes bug: 1183283
Change-Id: I23f05fe64475c3efe285e05a258c4625b801375c
Remove this sample config file from emc folder based on discussions
at the IRC meeting. Parameters specified in this sample file are
already documented in Cinder Admin Guide.
Change-Id: I4f742eedde962aefc6d5ab14d08a955859ae0e4f
The SolidFire device inherits all settings (including QoS) from
the parent on a clone. This meant that although we were expecting
QoS to be updated when a clone was issued with a different Volume
Type than the parent volume it was never acted upon.
This change fixes that bug by simply adding the *new* qos settings
to the ModifyVolume call after cloning.
Fixes bug: 1184235
Change-Id: I81786a07312027d543b9f06e54dfbf57e6085bc9
In function create_snapshot for cinder volume, '%' was missed
before variable snapshot from message defination. It was causing
an error during translation of debug message. Included '%'
for getting proper debug log.
Fixes: bug #1183736
Change-Id: I894639a4e5a69be80f4f53ec9d4f9f8648870173
Even though we're moving to testr and this will eventually
go away, we should maintain our ability to run and debug
individual tests until the broader change is released.
This just adds some text around the nosetests call
to get us back where we were.
Fixes bug: 1181996
Change-Id: Id0082ca8954664679430cc602425b1fadd29151b
If a user runs something like "cinder-manage volume delete a1234",
a ValueError is thrown because it fails to cast to int.
Catch this and treat the parameter as an id which will result in a
later VolumeNotFound error rather than breaking this way.
Change-Id: I95a9b9d7628cebe4b6d855ea925b0ad3a5f1c4c4
Clones copy over the attributes from the source volume,
we want to keep some admin data like source-volid and if
it's a clone or not. To do that we just need to call modifyvolume
after the clone operation and update this info.
Change-Id: I903fe6f8a57880abad8d315d091c1e57624a519c
When cloning a volume with SolidFire driver, the owner of
the cloned volume is not set correctly in SolidFire when there is
a transfer of ownership.
This results in inconsistent states where the cloned volume
is owned by the new tenant in Cinder but SolidFire thinks it is still
owned by the original volume's tenant.
This patch adds the newAccountID parameter to all CloneVolume calls.
If the cloned volume is owned by the same tenant, newAccountID will
be set to the same value as the original SolidFire volume. There
will be no change of ownership done by the cloning process in SolidFire.
If the cloned volume should be owned by a different tenant, newAccountID
will be set to the appropriate SolidFire account corresponding
to the new tenant. If the SolidFire account does not exist already,
it will be created.
Fixes: bug #1183521
Change-Id: I622ca2962478298e3e0c5a26866e39919805075f
Make the NFS share selection more intelligent by using a used_ratio
and oversub_ratio to determine the valid target. The used_ratio
is a hard-cap of maximum "actual" used space, oversub_ratio allows
for oversubscription of the share based upon the apparant-allocated
space for sparse files.
Removed Options: nfs_disk_util
New options added:
* nfs_used_ratio:
Float representation of the maximum allowed usage (e.g. 0.95
calculated by available_space / total_space) before the share
is no longer concidered a valid target. Default: 0.95
* nfs_oversub_ratio:
Float representation of the oversubscriotion ratio allowed when
utilizing sparse files (determined via "du"). This gets
multiplied against the total_capacity before comparing the
apparant space used by the backing volume files. Default
ratio is 1.0 (meaning reserved/allocated space cannot exceed
total space available on the mount).
DocImpact: New Configuration Options / Removal of Config Option
bp nfs-share-selection-logic-improvement
Change-Id: I3572e106eb4c794b08bf6f499e2da2d494c4072d
Since the fake filter scheduler does not update the capabilities, the sevice
states remains empty. To avoid the None value for free_capacity_gb, we set the
host_manager to a fake host manager with a list of service states.
Fixed Bug 1171906.
Change-Id: I359ea6952392cc5e6b8eff513f3ca4d8161a30e3
We've done this implementation because we need to use scheduler hint in
cinder with some specific filters. So, most part of code have been
imported from nova.
bp scheduler-hints
docimpact
Change-Id: I4c8a78ade4ff668d79e7aa6d0d358029754e3d90
The update of log was mostly to get coloring to work again
in devstack, and jsonutils was picked up as well.
The update to jsonutils also requires the six dependency
which is added to pip-requires.
Change-Id: I95d23aafe19b12d4a427149aa5f7a66394d178c6
Make sure VolumeNotFound is raised with the correct volume_id
in the SolidFire driver. Some exceptions were wrongfully referring
to an undefined variable named uuid.
Fixes: bug #1182966
Change-Id: Ied101fc9778f51dfb74b39f466c2c13d6e775a40
This patch adds support for mount options for NFS/GlusterFS volumes.
To set options, append them to the line for the volume in the
nfs_shares_config or glusterfs_shares_config file:
192.168.175.166:/testvol -o backupvolfile-server=192.168.175.177
host1:/testvol2 -o backupvolfile-server=host2,rw,other_option
These options are used when mounting from the cinder-volume
service and the nova-compute service. (There is an associated
nova-compute patch to enable the same.)
This is a useful change for GlusterFS because the
backupvolfile-server option enables a mount to succeed even if
the first server specified is offline.
Change-Id: I13bb0a930d6a05e803ccb50e6dc5d4b2b412ac11
The only functional difference here is that the value '2' is now
treated as False instead of True. I suspect that's ok.
Change-Id: I30e75470cb3c312d7f92d389995eb6d845dd2fb0
