| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
The issue has been resolved by infra [0], so we can use upper
constraints for all targets.
[0]
http://lists.openstack.org/pipermail/openstack-dev/2016-August/101474.html
Change-Id: I5c0378bdc6ddb4f8d0a328f3e61092be6cd579ac
(cherry picked from commit 6e775b82f4a9b82fdf72f82239d0e8d53add6ba2)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This will force pip install to use the upper-constraints.txt specified
version of pip modules. When you don't do this, you are out on the
bleeding edge and become unstable everytime some python library in the
wide world changes in a way that you don't expect.
Post jobs are not yet compatible with the upper-constraints file, so
override the install_command there to skip upper-constraints.
Blatantly cribbed from the equivalent trove patch:
https://review.openstack.org/#/c/298850
Change-Id: I73aa8cf37adc36db2c56a8f29ea165775a83dc7f
Closes-Bug: #1563038
(cherry picked from commit e27e7ed4a9452eb69f6c1af434adb5620706c227)
|
|
|
|
| |
Change-Id: I55cb9b4472b759ed0fa3ae0575e7c4aee43cfe55
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We currently mask passwords in driver_info for all API responses,
except for agent lookups. This is because driver_vendor_passthru
just expects a dict to return as JSON in the response, and doesn't
modify it at all.
Change lookup to follow the defined policy when returning the node
object in the response, the same way we do for other API responses.
Co-authored-by: Jim Rollenhagen <jim@jimrollenhagen.com>
Change-Id: Ib19d2f86d3527333e905fdbf1f09fc3dc8b8c5a7
Closes-Bug: #1572796
(cherry picked from commit 426a306fb580762e97ada04e1253dedd9b64d410)
|
|
|
|
| |
Change-Id: I5dca21edc38944fb707a795041dba7e801c3037d
|
|
|
|
|
| |
Change-Id: Ia0c469669417f7de7b35ccf517750b9ecaebbed0
Closes-Bug: #1583535
(cherry picked from commit 4f463617ac9274679185993f058741bd231d9dfc)
|
|\ |
|
| |
| |
| |
| | |
Change-Id: I3c324417492e211d0924198c19e4a9fb058bd4b7
|
|/
|
|
|
|
|
|
|
|
| |
This patch forces iRMC vmedia boot from remotely connected
(redirected) CD/DVD using 'ipmitool raw' command instead of 'ipmitool
chassis bootdev cdrom'.
Closes-Bug: #1561852
Change-Id: I606e4d3a630ddc4eed071773afcb5274ee64a439
(cherry picked from commit 03fe93abfe5779f27c0eee5495f189e0104a5184)
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Automated cleaning fails for iLO driver if iLO is in System POST
state. iLO does not allow setting of boot device when it is in
System POST state. Node needs to be powered off before initiating
boot for cleaning.
(cherry-picked from commit 38201d9f2adff4cae77578445a528cdded9bcc87)
This change applies only to agent_ilo driver that supports in-band
cleaning in Liberty release.
Fix is made in IloVirtualMediaAgentDeploy:prepare_cleaning().
Closes-Bug: 1536730
Change-Id: Icd0f39d718aabdcad56636e662dec788c4e1b427
|
|\ \
| | |
| | |
| | | |
stable/liberty
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Disable default execution of clean step 'reset_ilo' during automated
cleaning. Resetting of iLO is not required during every invocation
of automated cleaning.
(cherry picked from commit 380d6bfc52a793759f25955e4ac155893f427cb4)
Change-Id: I45a5542c8db339417916c64455eba0fae9c3f44a
Closes-Bug: 1542977
|
|/
|
|
|
|
|
|
|
|
|
|
|
| |
Two unittests in the test_glance_service.py module was depending on
getting an instance of the real glanceclient and it was failing due a
new glanceclient release. This patch is fixing the problem by mocking
the client instanciation.
Closes-bug: #1552839
Closes-Bug: #1552834
Co-Authored-By: Lucas Alvares Gomes <lucasagomes@gmail.com>
Change-Id: Ic9dc3f1016667c6d7ec5ee324d5fff48af82c581
(cherry picked from commit 036b959b9245ba44b5ea895ea48584dedc282fdf)
|
|\
| |
| |
| | |
stable/liberty
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In order to the pxe_drac to work with IPA ramdisks, the vendor interface
should expose the lookup() method. This patch is adding it to the list
of mapped vendor methods for that driver.
(cherry picked from commit 3afbfaf40a80972d824f0aed7bb649a27755b323)
Closes-Bug: #1541858
Change-Id: I0cc13550bbf980ee48579d750120d4c33d1fe6a1
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In Ia575ba803a0fb70f39146bd75d381ed19414bd23, oslo.context
added roles support in the context itself. Once that change
is released in oslo.context and the global requirements has
been updated, we should switch to passing in the roles in
the __init__ parameter. Until then we should set
self.roles *after* the constructor since the constructor
sets the roles to None when the new library gets released.
Closes-Bug: 1549317
Change-Id: Ie28a4144ccac5d6894405ba7f801617376e35c51
(cherry picked from commit b0ae5e9f2dc078af8e80a284c8692b79e98e0ed2)
|
|\ \
| | |
| | |
| | | |
stable/liberty
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When cleaning is abort Ironic leaves the node's clean_step field
populated with the step that was running at the moment. Because of that,
once the next cleaning starts again (after the node have been moved to
"manageable" and in the transition to "available") it will hang and
never finish, the reason for that is because once the node starts
heartbeating again Ironic won't know if the cleaning just started or
it's just continuing cleaning (it uses clean_step as a flag for it,
see the bug linked with this patch).
This patch fix this by cleaning up the node's clean_field once the
cleaning process is aborted.
Closes-Bug: #1518374
Change-Id: I423e89f9574a8952875ad9f03ce33298cb5660e7
(cherry picked from commit a8070a59233895ce1c7b6f92ef88cc31bc729188)
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The goto :boot_whole_disk section in the iPXE template is bogus, the
"kernel" command is being used to chain the "chain.c32" file, that
command should be "chain" and the "append" command is not a valid iPXE
command.
This patch replaces those commands with the "sanboot" command from iPXE
which is (among other things) used to chain the boot process to the
local disk.
Closes-Bug: #1524403
Change-Id: I7d2310d94e40b1c9c76957382ab19c0076bf42cd
(cherry picked from commit 952d5c5a06f35181768e52dd660afe548bc6fa60)
|
| |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently, BaseAgentVendor.hearbeat handles uncaught exceptions by
advancing FSM with the 'failed' event no matter the state. Having
encountered an error, some drivers may both advance FSM themselves and
raise an exception. This leads to advancing the FSM with the 'fail' event
while already in the DEPLOYFAIL state.
Suggested patch purpose is to advance the FSM only if either in DEPLOYWAIT
or DEPLOYING state.
Change-Id: Ie2498bcb095535a25e46cea78365c2bfa4562d66
Closes-Bug: #1506376
(cherry picked from commit c6c885858a658b0a2d11be200133a3f530f1b6a1)
|
| |
| |
| |
| |
| |
| |
| | |
For more information about this automatic import see:
https://wiki.openstack.org/wiki/Translations/Infrastructure
Change-Id: I4fbb5efbedab5d53f97ed5a0c5fa86c884b9cb21
|
|/
|
|
|
|
|
|
|
|
|
|
| |
When an IPMI password consists only of numbers, Ironic fails to create a
password file. From the format point of view, these passwords are
correct. These are that it is stored to the variable after being
converted to an int type, cause errors in the subsequent processing to
expect that it is a string type.
Change-Id: I85ee2e904fdb0d538f69c7dbfab4c81bb080b8cf
Closes-Bug: #1489234
(cherry picked from commit 93f99381acd727f539d3a78c0ae82b95860ba649)
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
pecan depends on WebTest 1.3.1, but we use JSON features from 2.0.
Closes-Bug: #1524816
Change-Id: I9b99c686cddc4a916c39359d5a1006ed4c1ddef0
(cherry picked from commit 40f9238275e12d2d9bad97c20147dd451f841125)
|
|\ \ |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| | |
This patch updates python-scciclient version number for iRMC drivers.
Change-Id: I3f0a15b09d05379f6e4e59c81754c2c76712200b
closes-bug: #1518999
closes-bug: #1519000
(cherry picked from commit e936fbfda1294c050674332a204ce4ee77aa1f8e)
|
|\ \ |
|
| | |
| | |
| | |
| | | |
Change-Id: I08c55bd02c14a165a160700ba5389576f0e666c4
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
There was a bug discovered in ImcSdk 0.7.1 with the SSL handshake that
caused problems when running Ironic for longer periods of time, this
patch changes the required version to the latest version of the ImcSdk
which contains a fix for this issue.
Change-Id: I7665e953b73ed83d5fb5d8adfbed5b3647e5ed48
Closes-Bug: #1508849
(cherry picked from commit 409c603761c0f1d5a5c3113e22c86cbcdbb5cc3b)
|
|\ \ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
When using web server for hosting images for ilo driver,
take_over should recreate the boot ISO, if it was created
by Ironic. If the boot iso is not recreated, then the
node doesn't boot up successfully when a node is powered off
and powered on.
The fix is available in master branch via
I5481e705892f66cb3184ac16ff56c123f5c69087 in file boot.py
from lines 117 to 122. The test case for the fix is available via
https://review.openstack.org/#/c/239627 in master branch.
The master branch patch I5481e705892f66cb3184ac16ff56c123f5c69087
implements a feature (~2K lines of change) while the fix for this
issue is just 5-6 lines of change.
Closes-bug: #1510425
Change-Id: I6b7f3901f773a5aba906c50dcbd574412f7b6af3
|
|\ \ \ \ \
| |_|_|/ /
|/| | | |
| | | | | |
into stable/liberty
|
| | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The ${mac} variable not necessarily points to the MAC address of the NIC
we are booting from. This patch updates the boot.ipxe script to make it
more robust by interating over all NICs found in the machine and boot
from the first one capable of booting (the one that contains a
configuration file in the server). If no NICs are capable of booting the
script will fail just like before telling the user that the network boot
failed because we couldn't find any configuration for the present NICs.
In case the inc command is not found, the boot.ipxe will fall back to
the old behavior.
This backport is the combination of 2 commits (squashed):
(cherry picked from commit 1c28eb79a5d1b25fb685e2f48449c51e905766b5)
(cherry picked from commit f968e1a22d3ed928278a04f0da9124678474ff0a)
We need to squash both patches because they must merge at the same time
to avoid breakages.
Closes-Bug: #1507738
Closes-Bug: #1504482
Change-Id: I5b8f0740fa42f611e84b66dc7f6578cffe02760a
|
|/ / /
| | |
| | |
| | | |
Change-Id: Ib15f2161e3e14afbc496bcc21b326cda21a0c918
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
A bug was introduced during Liberty where Ironic transparently
ignores all clean steps and finishes cleaning. This is caused
by _get_node_next_clean_steps returning an empty list when
cleaning has just started. Fix this method to return the full
list of clean steps when cleaning begins.
This may leave previous tenants' data on disk and available to future
tenants. Deployers should apply this patch (or upgrade to a new release
with this patch) ASAP.
Closes-Bug: #1517277
(cherry picked from commit 1c700e6d62ad299e3fc9023e30b98d51408e49e1)
Change-Id: If815f81d7e668244f0d434d4e2933e8f41946107
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Mostly cargo-culted from I9549e6b491b26232f4dac900b0bf05f10eb0e30a.
Note that reno automatically inserts version numbers; liberty.rst is
sufficient for intermediate release models.
This is backported to stable/liberty by request in
http://lists.openstack.org/pipermail/openstack-dev/2015-November/078940.html
Change-Id: Ic700d719242b17532b0b43a68b7eff507cb3a43c
(cherry picked from commit aae12e5109dcff35a72f111d9ef5d72157d589c6)
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
agent_ilo driver doesn't cleanup the temporary
files created. This commit makes sure
that the temporary files are cleaned up before
the server goes for second boot.
Closes-bug: #1503795
Change-Id: I6cce1d1f3a5631e1af992c173379927fed3c55a1
(cherry picked from commit e34c4248af5f453a776af8b3933c5d7ece1c7b9c)
|
| |/ /
|/| |
| | |
| | |
| | |
| | |
| | | |
For more information about this automatic import see:
https://wiki.openstack.org/wiki/Translations/Infrastructure
Change-Id: Ie12fc317fb275a68d5f6f93bb8dc9c1fe15fee6a
|
| | |
| | |
| | |
| | | |
Change-Id: I126122de126b3f1095621d8b92709193b152e9e5
|
| |/
|/|
| |
| |
| |
| |
| |
| | |
This patch adds the defaultbranch tag to the .gitreview file for
stable/liberty to ensure that any reviews done from this branch end up
in the right place on gerrit.
Change-Id: I55db3f427b35cdc8e27a0fd45c5b6960f8becd97
|
|/
|
|
| |
Change-Id: I88c1cebe1d89b89f66bd92de7a6b29683381a481
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Not only 'dd' command but also 'parted' makes iscsi deploy fail in
Japanese environment. This patch applies the same fix as 'dd'.
ironic.drivers.modules.iscsi_deploy [-] Deploy failed for instance
<uuid>. Error: 'ascii' codec can't decode byte 0xe8 in position 20:
ordinal not in range(128)
Change-Id: If6c3fbd08ff380470d88b16605b44eb00c271352
Related-Bug: #1488289
(cherry picked from commit da5ddf4e703b6de9ef67420a0eb4baaa740980d7)
|
|
|
|
|
|
|
|
| |
oslo.db fixtures need these 2 packages and is breaking us.
Closes-Bug: #1503501
Change-Id: I75757b5bd7d9e75ca37d1b9a1efff810ba84b5df
(cherry picked from commit e7e0a056b4e2060ab53f2402c837a55e440edaaf)
|
|
|
|
| |
Change-Id: Ia4915c4d2ddc72f92db3bccb147ab45d2fc9957f
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When executing 'dd' command, baremetal provisioning fails on OS with
Japanese because of ascii codec error:
InstanceDeployFailure: Deploy failed for instance <uuid>.
Error: 'ascii' codec can't decode byte 0xc2 in position 24:
ordinal not in range(128)
This patch set fixes this bug by using standard locale when
executing 'dd'.
Change-Id: I8107975044cc9dec85e0d7571f99eaf5ea869891
Closes-Bug:#1488289
Co-Authored-By: Ruby Loo <rloo@yahoo-inc.com>
(cherry picked from commit 8166c6e8ddd333ee9930d0d7d1afb7c2ee47fc4a)
|
|\ \ |
|