- Kimchi-devel - oVirt List Archives

[PATCH 0/4] improve controller
by shaohef＠linux.vnet.ibm.com 21 Jan '14

21 Jan '14

From: ShaoHe Feng <shaohef(a)linux.vnet.ibm.com> add a method to load root sub collections/resouces automatically ShaoHe Feng (4): improve controller: add a method to load root sub collections/resouces automatically improve controller: tag the collections/resouces of root with @urlSubNode improve controller: Root loads collections/resouces automatically improve controller: set authentication automatically src/kimchi/control/__init__.py | 8 ++++++++ src/kimchi/control/config.py | 2 ++ src/kimchi/control/debugreports.py | 2 ++ src/kimchi/control/host.py | 2 ++ src/kimchi/control/interfaces.py | 2 ++ src/kimchi/control/networks.py | 2 ++ src/kimchi/control/plugins.py | 2 ++ src/kimchi/control/storagepools.py | 2 ++ src/kimchi/control/tasks.py | 2 ++ src/kimchi/control/templates.py | 2 ++ src/kimchi/control/vms.py | 2 ++ src/kimchi/root.py | 26 +++++-------------------- src/kimchi/server.py | 12 +++++------- src/kimchi/utils.py | 39 ++++++++++++++++++++++++++++++++++++++ 14 files changed, 77 insertions(+), 28 deletions(-) -- 1.8.4.2

4 12

debugreport generation error
by Christy Perez 21 Jan '14

21 Jan '14

Hi, I randomly decided to try out the debugreport feature on my Fedora 20 latptop today and ran into an issue. The first odd thing was that I was asked to login to the kimchi portal again, but maybe that's on purpose? Next, the action seems to hang indefinitely. I scrolled back up and found this call trace: 127.0.0.1 - christy [20/Jan/2014:15:23:45] "GET /tasks/1 HTTP/1.1" 200 54 "https://localhost:8001/" "Mozilla/5.0 (X11; Linux x86_64; rv:26.0) Gecko/20100101 Firefox/26.0" gluster Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/sos/sosreport.py", line 937, in setup plug.setup() File "/usr/lib/python2.7/site-packages/sos/plugins/gluster.py", line 91, in setup gluster_major = int((pkg["version"])[:1]) TypeError: string indices must be integers, not str 127.0.0.1 - christy [20/Jan/2014:15:23:45] "GET /tasks/1 HTTP/1.1" 200 54 "https://localhost:8001/" "Mozilla/5.0 (X11; Linux x86_64; rv:26.0) Gecko/20100101 Firefox/26.0" yum Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/sos/sosreport.py", line 937, in setup plug.setup() File "/usr/lib/python2.7/site-packages/sos/plugins/yum.py", line 29, in setup rhelver = self.policy().rhel_version() AttributeError: 'FedoraPolicy' object has no attribute 'rhel_version' I did find that someone else is hitting this issue: https://bugzilla.redhat.com/show_bug.cgi?id=1028102 (which is actually from an IBM bug...) https://bugzilla.linux.ibm.com/show_bug.cgi?id=99234 And I also found this opened against kimchi: https://bugzilla.linux.ibm.com/show_bug.cgi?id=101135 So I tried running it myself in a terminal, and I got an error: $ sosreport sosreport (version 3.0) no valid plugins were enabled So I tried running it as root: $ sudo sosreport [sudo] password for christy: sosreport (version 3.0) This command will collect system configuration and diagnostic information from this Fedora system. An archive containing the collected information will be generated in /var/tmp. For more information on the Fedora Project visit: https://fedoraproject.org/ The generated archive may contain data considered sensitive and its content should be reviewed by the originating organization before being passed to any third party. No changes will be made to system configuration. Press ENTER to continue, or CTRL-C to quit. Please enter your first initial and last name [christy-toshiba]: cperezPlease enter the case number that you are generating this report for: gluster Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/sos/sosreport.py", line 937, in setup plug.setup() File "/usr/lib/python2.7/site-packages/sos/plugins/gluster.py", line 91, in setup gluster_major = int((pkg["version"])[:1]) TypeError: string indices must be integers, not str yum Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/sos/sosreport.py", line 937, in setup plug.setup() File "/usr/lib/python2.7/site-packages/sos/plugins/yum.py", line 29, in setup rhelver = self.policy().rhel_version() AttributeError: 'FedoraPolicy' object has no attribute 'rhel_version' Running plugins. Please wait ... Running 39/65: networking... So this time it worked better, but, not completely. It hung again there at 39/65. I eventually got an sosreport (including the above traceback still) by using: $ sudo sosreport -v --batch -n networking -n systemd Running plugins. Please wait ... Running 25/63: kernel... caught IO error copying /sys/module/kdb_main/parameters/enable_nmi Running 46/63: processor... caught IO error copying /sys/devices/system/cpu/cpu0/power/autosuspend_delay_ms caught IO error copying /sys/devices/system/cpu/cpu1/power/autosuspend_delay_ms caught IO error copying /sys/devices/system/cpu/cpu2/power/autosuspend_delay_ms caught IO error copying /sys/devices/system/cpu/cpu3/power/autosuspend_delay_ms caught IO error copying /sys/devices/system/cpu/power/autosuspend_delay_ms Running 63/63: yum... Creating compressed archive... Your sosreport has been generated and saved in: /var/tmp/sosreport-christy-toshiba-20140120155033.tar.xz The checksum is: dd42f749b1d6ae0952a71bd6ce12b80c Please send this file to your support representative. === I just wanted to share my findings, in case there are more bugs opened for this. It seems to be a problem for Power and x86. Regards, - Christy

2 1

Re: [Kimchi-devel] [project-kimchi] [PATCH V3 0/5] Host's package management support
by Hongliang Wang 21 Jan '14

21 Jan '14

Hi Paulo, I'm working on the UI for package management, though this patch set seems needs rebased. Could send a new one? On 12/07/2013 04:19 AM, Paulo Vital wrote: > V3 -> V2: > Fixed API.json file > Modified _repos_storage type to dictionary, and adjust methods to use it > Changed return of some functions to raise exceptions instead of boolean > > V1 -> V2: > Cleaned whitespaces and fixed some typos > Changed return of backend functions to raise exceptions instead of boolean > > V1: > > This patch set provides support to host's package management operations. > At this point, an agnostic class is providing support to backend and > REST API operations. > > In backend, YUM support to operate with repositories is provided by this > patch set. The support to APT repositories will be provided in a close > future. > > To test the backend execute the following commands: > $ sudo tests/run_tests.sh test_model.ModelTests.test_repository_create > $ sudo tests/run_tests.sh test_model.ModelTests.test_repository_update > $ sudo tests/run_tests.sh test_model.ModelTests.test_repository_disable_enable > > To test the REST API, execute the following commands (all them are agnostic of the > host's distro): > > 1) Get list of all repositories enabled in the host: > $curl -H 'Content-type: application/json' -H 'Accept: application/json' http://localhost:8000/repositories/ -X GET > > 2) Create a new repository: > $ curl -H 'Content-type: application/json' -H 'Accept: application/json' http://localhost:8000/repositories/ -X POST -d' > { > "repo_id":"fedora-fake", > "repo_name":"Fedora 19 FAKE", > "baseurl": "http://www.fedora.org" > } > ' > > 3) Get information from a specific repository: > $curl -H 'Content-type: application/json' -H 'Accept: application/json' http://localhost:8000/repositories/fedora-fake > > 4) Update a specific repository: > $ curl -H 'Content-type: application/json' -H 'Accept: application/json' http://localhost:8000/repositories/fedora-fake -X PUT -d' > { > "repo_id":"fedora-fake", > "repo_name":"Fedora 19 FAKEs", > "baseurl": "http://www.fedora.org/downloads" > } > ' > > 5) Disable a specific repository: > $ curl -H 'Content-type: application/json' -H 'Accept: application/json' http://localhost:8000/repositories/fedora-fake/disable -X POST -d '' > > 6) Enable a specific repository: > $ curl -H 'Content-type: application/json' -H 'Accept: application/json' http://localhost:8000/repositories/fedora-fake/enable -X POST -d '' > > 6) Delete a specific repository: > $ curl -H 'Content-type: application/json' -H 'Accept: application/json' http://localhost:8000/repositories/fedora-fake -X DELETE > > > Paulo Vital (5): > Host's package management: Update API.md > Host's package management: Update REST API > Host's package management: Update backend. > Host's package management: Update Makefile > Host's package management: Update test-cases > > docs/API.md | 90 ++++++++++ > src/kimchi/API.json | 65 +++++++ > src/kimchi/Makefile.am | 1 + > src/kimchi/controller.py | 20 +++ > src/kimchi/mockmodel.py | 33 ++++ > src/kimchi/model.py | 33 ++++ > src/kimchi/repositories.py | 419 +++++++++++++++++++++++++++++++++++++++++++++ > src/kimchi/server.py | 1 + > tests/test_model.py | 99 +++++++++++ > tests/test_rest.py | 32 ++++ > 10 files changed, 793 insertions(+) > create mode 100644 src/kimchi/repositories.py >

2 3

[PATCH v2] RollbackContext: Make it complied with Python Standard
by Zhou Zheng Sheng 21 Jan '14

21 Jan '14

According to Python documentation, context managers should not re-raise the exception passed to __exit__ method. If __exit__() does not return True, Python re-raises this exception automatically. Context managers should only raise exception from __exit__() itself. This means we should only re-raise exception caught from the undo functions. Another problem of the current implementation is that it suppresses str exceptions. This is because of bug in old Python version. When an exception is raised from within a "with" statement, exc_value would be a raw str or int rather than Exception instance, so when we re-raise the stored exc_value, it triggers another exception saying "exceptions must be old-style classes or derived from BaseException, not str". This patch fix this problem by raising the stored exception value along with its store exception type. This patch also adds unit test for RollbackContext. http://docs.python.org/2/library/stdtypes.html#contextmanager.__exit__ v1 -> v2 Add tests/test_rollbackcontext.py to tests/Makefile.am. Add tests/test_rollbackcontext.py to Makefile.am PEP8 list. Signed-off-by: Zhou Zheng Sheng <zhshzhou(a)linux.vnet.ibm.com> --- Makefile.am | 1 + src/kimchi/rollbackcontext.py | 26 +++++------ tests/Makefile.am | 1 + tests/test_rollbackcontext.py | 102 ++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 117 insertions(+), 13 deletions(-) create mode 100644 tests/test_rollbackcontext.py diff --git a/Makefile.am b/Makefile.am index 04ad696..b09a545 100644 --- a/Makefile.am +++ b/Makefile.am @@ -60,6 +60,7 @@ PEP8_WHITELIST = \ tests/test_model.py \ tests/test_plugin.py \ tests/test_rest.py \ + tests/test_rollbackcontext.py \ tests/test_storagepool.py \ tests/utils.py \ $(NULL) diff --git a/src/kimchi/rollbackcontext.py b/src/kimchi/rollbackcontext.py index 2afd114..d493af2 100644 --- a/src/kimchi/rollbackcontext.py +++ b/src/kimchi/rollbackcontext.py @@ -6,6 +6,7 @@ # Authors: # Adam Litke <agl(a)linux.vnet.ibm.com> # ShaoHe Feng <shaohef(a)linux.vnet.ibm.com> +# Zhou Zheng Sheng <zhshzhou(a)linux.vnet.ibm.com> # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public @@ -45,24 +46,23 @@ class RollbackContext(object): return self def __exit__(self, exc_type, exc_value, traceback): - firstException = exc_value - + """ + According to Python official doc. This function should only re-raise + the exception from undo functions. Python automatically re-raises the + original exception when this function does not return True. + http://docs.python.org/2/library/stdtypes.html#contextmanager.__exit__ + """ + undoExcInfo = None for undo, args, kwargs in self._finally: try: undo(*args, **kwargs) - except Exception as e: + except Exception: # keep the earliest exception info - if not firstException: - firstException = e - # keep the original traceback info - traceback = sys.exc_info()[2] + if undoExcInfo is None: + undoExcInfo = sys.exc_info() - # re-raise the earliest exception - if firstException is not None: - if type(firstException) is str: - sys.stderr.write(firstException) - else: - raise firstException, None, traceback + if exc_type is None and undoExcInfo is not None: + raise undoExcInfo[0], undoExcInfo[1], undoExcInfo[2] def defer(self, func, *args, **kwargs): self._finally.append((func, args, kwargs)) diff --git a/tests/Makefile.am b/tests/Makefile.am index 55d3ab4..59e344d 100644 --- a/tests/Makefile.am +++ b/tests/Makefile.am @@ -31,6 +31,7 @@ EXTRA_DIST = \ test_osinfo.py \ test_plugin.py \ test_rest.py \ + test_rollbackcontext.py \ test_server.py \ test_storagepool.py \ test_vmtemplate.py \ diff --git a/tests/test_rollbackcontext.py b/tests/test_rollbackcontext.py new file mode 100644 index 0000000..fea430d --- /dev/null +++ b/tests/test_rollbackcontext.py @@ -0,0 +1,102 @@ +# +# Project Kimchi +# +# Copyright IBM, Corp. 2014 +# +# Authors: +# Zhou Zheng Sheng <zhshzhou(a)linux.vnet.ibm.com> +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library; if not, write to the Free Software +# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + +import unittest + +from kimchi.rollbackcontext import RollbackContext + + +class FirstError(Exception): + '''A hypothetical exception to be raise in the test firstly.''' + pass + + +class SecondError(Exception): + '''A hypothetical exception to be raise in the test secondly.''' + pass + + +class RollbackContextTests(unittest.TestCase): + + def setUp(self): + self._counter = 0 + + def _inc_counter(self): + self._counter += 1 + + def _raise(self, exception=FirstError): + raise exception() + + def test_rollback(self): + with RollbackContext() as rollback: + rollback.prependDefer(self._inc_counter) + rollback.prependDefer(self._inc_counter) + self.assertEquals(self._counter, 2) + + def test_raise(self): + try: + with RollbackContext() as rollback: + rollback.prependDefer(self._inc_counter) + rollback.prependDefer(self._inc_counter) + raise FirstError() + rollback.prependDefer(self._inc_counter) + except FirstError: + # All undo before the FirstError should be run + self.assertEquals(self._counter, 2) + else: + self.fail('Should have raised FirstError') + + def test_raise_undo(self): + try: + with RollbackContext() as rollback: + rollback.prependDefer(self._inc_counter) + rollback.prependDefer(self._raise) + rollback.prependDefer(self._inc_counter) + except FirstError: + # All undo should be run + self.assertEquals(self._counter, 2) + else: + self.fail('Should have raised FirstError') + + def test_raise_prefer_original(self): + try: + with RollbackContext() as rollback: + rollback.prependDefer(self._raise, SecondError) + raise FirstError() + except FirstError: + pass + except SecondError: + self.fail('Should have preferred FirstError to SecondError') + else: + self.fail('Should have raised FirstError') + + def test_raise_prefer_first_undo(self): + try: + with RollbackContext() as rollback: + rollback.prependDefer(self._raise, SecondError) + rollback.prependDefer(self._raise, FirstError) + except FirstError: + pass + except SecondError: + self.fail('Should have preferred FirstError to SecondError') + else: + self.fail('Should have raised FirstError') -- 1.7.11.7

4 3

[PATCH] RollbackContext: Make it complied with Python Standard
by Zhou Zheng Sheng 21 Jan '14

21 Jan '14

According to Python documentation, context managers should not re-raise the exception passed to __exit__ method. If __exit__() does not return True, Python re-raises this exception automatically. Context managers should only raise exception from __exit__() itself. This means we should only re-raise exception caught from the undo functions. Another problem of the current implementation is that it suppresses str exceptions. This is because of bug in old Python version. When an exception is raised from within a "with" statement, exc_value would be a raw str or int rather than Exception instance, so when we re-raise the stored exc_value, it triggers another exception saying "exceptions must be old-style classes or derived from BaseException, not str". This patch fix this problem by raising the stored exception value along with its store exception type. http://docs.python.org/2/library/stdtypes.html#contextmanager.__exit__ Signed-off-by: Zhou Zheng Sheng <zhshzhou(a)linux.vnet.ibm.com> --- src/kimchi/rollbackcontext.py | 25 +++++------ tests/test_rollbackcontext.py | 102 ++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 114 insertions(+), 13 deletions(-) create mode 100644 tests/test_rollbackcontext.py diff --git a/src/kimchi/rollbackcontext.py b/src/kimchi/rollbackcontext.py index 2afd114..9502169 100644 --- a/src/kimchi/rollbackcontext.py +++ b/src/kimchi/rollbackcontext.py @@ -45,24 +45,23 @@ class RollbackContext(object): return self def __exit__(self, exc_type, exc_value, traceback): - firstException = exc_value - + """ + According to Python official doc. This function should only re-raise + the exception from undo functions. Python automatically re-raises the + original exception when this function does not return True. + http://docs.python.org/2/library/stdtypes.html#contextmanager.__exit__ + """ + undoExcInfo = None for undo, args, kwargs in self._finally: try: undo(*args, **kwargs) - except Exception as e: + except Exception: # keep the earliest exception info - if not firstException: - firstException = e - # keep the original traceback info - traceback = sys.exc_info()[2] + if undoExcInfo is None: + undoExcInfo = sys.exc_info() - # re-raise the earliest exception - if firstException is not None: - if type(firstException) is str: - sys.stderr.write(firstException) - else: - raise firstException, None, traceback + if exc_type is None and undoExcInfo is not None: + raise undoExcInfo[0], undoExcInfo[1], undoExcInfo[2] def defer(self, func, *args, **kwargs): self._finally.append((func, args, kwargs)) diff --git a/tests/test_rollbackcontext.py b/tests/test_rollbackcontext.py new file mode 100644 index 0000000..23b41c4 --- /dev/null +++ b/tests/test_rollbackcontext.py @@ -0,0 +1,102 @@ +# +# Project Kimchi +# +# Copyright IBM, Corp. 2013 +# +# Authors: +# Zhou Zheng Sheng <zhshzhou(a)linux.vnet.ibm.com> +# +# This library is free software; you can redistribute it and/or +# modify it under the terms of the GNU Lesser General Public +# License as published by the Free Software Foundation; either +# version 2.1 of the License, or (at your option) any later version. +# +# This library is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +# Lesser General Public License for more details. +# +# You should have received a copy of the GNU Lesser General Public +# License along with this library; if not, write to the Free Software +# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + +import unittest + +from kimchi.rollbackcontext import RollbackContext + + +class FirstError(Exception): + '''A hypothetical exception to be raise in the test firstly.''' + pass + + +class SecondError(Exception): + '''A hypothetical exception to be raise in the test secondly.''' + pass + + +class RollbackContextTests(unittest.TestCase): + + def setUp(self): + self._counter = 0 + + def _inc_counter(self): + self._counter += 1 + + def _raise(self, exception=FirstError): + raise exception() + + def test_rollback(self): + with RollbackContext() as rollback: + rollback.prependDefer(self._inc_counter) + rollback.prependDefer(self._inc_counter) + self.assertEquals(self._counter, 2) + + def test_raise(self): + try: + with RollbackContext() as rollback: + rollback.prependDefer(self._inc_counter) + rollback.prependDefer(self._inc_counter) + raise FirstError() + rollback.prependDefer(self._inc_counter) + except FirstError: + # All undo before the FirstError should be run + self.assertEquals(self._counter, 2) + else: + self.fail('Should have raised FirstError') + + def test_raise_undo(self): + try: + with RollbackContext() as rollback: + rollback.prependDefer(self._inc_counter) + rollback.prependDefer(self._raise) + rollback.prependDefer(self._inc_counter) + except FirstError: + # All undo should be run + self.assertEquals(self._counter, 2) + else: + self.fail('Should have raised FirstError') + + def test_raise_prefer_original(self): + try: + with RollbackContext() as rollback: + rollback.prependDefer(self._raise, SecondError) + raise FirstError() + except FirstError: + pass + except SecondError: + self.fail('Should have preferred FirstError to SecondError') + else: + self.fail('Should have raised FirstError') + + def test_raise_prefer_first_undo(self): + try: + with RollbackContext() as rollback: + rollback.prependDefer(self._raise, SecondError) + rollback.prependDefer(self._raise, FirstError) + except FirstError: + pass + except SecondError: + self.fail('Should have preferred FirstError to SecondError') + else: + self.fail('Should have raised FirstError') -- 1.7.11.7

2 6

[PATCH] Abstract vm element class for live update
by lvroyce0210＠gmail.com 20 Jan '14

20 Jan '14

From: Royce Lv <lvroyce(a)linux.vnet.ibm.com> Abstract vm element class for live update to be compatible with different element type and their various libvirt calls. Royce Lv (1): Abstract vm element updater for common usage src/kimchi/model.py | 65 +++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 63 insertions(+), 2 deletions(-) -- 1.8.1.2

3 3

[PATCH 0/3] (WIP) Storagepool SCSI/FC
by Rodrigo Trujillo 20 Jan '14

20 Jan '14

NOTICE: THIS FEATURE IS STILL UNDER DISCUSSION/AGREEMENT/DEVELOPMENT This patch set implements the basic functions in order to create a SCSI storagepool based in fibre channel protocol. --- What is implemented? --- * backend API It is possible to create a pool with curl POST request like: curl -X POST -u <USER> -H 'Content-type: application/json' \ -H 'Accept: application/json' http://localhost:8000/storagepools \ --data '{ \ "name": "testFC", \ "source": { "adapter_type": "scsi_host", \ "adapter_name": "scsi_hostX", \ "wwnn":"112233789", \ "wwpn":"44332211" }, \ "path": "/dev/disk/by-id", \ "type": "scsi" }' * json schema updated * created a function to test the host libvirt version SCSI FC pool and volumes have theirself configuration only after libvirt version 1.0.5. If lesser, I must use older configuration tags * updates the VM guest template with proper disk xml Create volumes in a SCSI pool is not supported by libvirt, because volumes are the LUNS preconfigured and available in the SAN. Disk will have the biggest LUN found assigned. --- What does need to be discussed? Ideas? --- 1) Backend expects right information from UI, so UI would have to ask and send 'adapter_name', 'wwnn' and 'wwpn' (other fields can be fi xed). However, the UI coud show the FC_Hosts available and let user select which one he wants. This would require to implement more ba ck-end functions to gather FC_host info, automatically (name, wwnn, wwpn). So, at this moment, just create an UI with input fields or facili tate the users life ? 2) Currently the user have no choice on with LUN to assign to a guest. LUN is selected automatically and if not LUN is available, guest will not have any disk. In order to improve this, the guest creation fase could have one more step, if the pool in the template in SCSI type. If SCSI, user could choose which LUN he wants to assign to VM. Notice that the disk size in the template have been, and will be, ig nored. What do you guys think about this approach ? 3) LUNs refresh: Once the storage admin creates more LUNs, the host admin needs to perform some commands in order to kernel 'see' and map new devices. I am implementing this backend functionality. A refresh bu tton will be necessary in the UI, where could it be placed? - In the storagepool creation window ? When select SCSI type - In the window to select the LUN, when creating new guest (if approved this extra step) - In the storagepool tab, in the pool section, if its a SCSI type - Also, when the user expands a SCSI pool section, which will show the volumes (LUNs), this function can be triggered. 4) More ideas/concerns/suggestions ? Rodrigo Trujillo (3): Storagepool: SCSI/Fibre Channel backend implementation Storagepools: Function to check libvirt version Storagepools: Assign SCSI fibre channel LUN as disk to a new guest docs/API.md | 6 ++++ src/kimchi/API.json | 18 ++++++++++- src/kimchi/model.py | 79 +++++++++++++++++++++++++++++++++++++++++++++--- src/kimchi/utils.py | 14 +++++++++ src/kimchi/vmtemplate.py | 40 +++++++++++++++++++++++- 5 files changed, 151 insertions(+), 6 deletions(-) -- 1.8.1.4

4 7

[PATCH V3] UI bug fix: Set vlan_id value only if 'Enable VLAN' is checked
by Aline Manera 20 Jan '14

20 Jan '14

From: Aline Manera <alinefm(a)br.ibm.com> V2 -> V3: - Add validation to vlan id - Reset vlan id input box when checkbox isn't checked - Update po files V1 -> V2: - Remove alert used for testing Aline Manera (1): UI bug fix: Set vlan_id value only if 'Enable VLAN' is checked po/en_US.po | 10 ++++++++-- po/kimchi.pot | 9 ++++++++- po/pt_BR.po | 10 ++++++++-- po/zh_CN.po | 10 ++++++++-- ui/js/src/kimchi.network.js | 14 +++++++++++--- ui/pages/i18n.html.tmpl | 1 + 6 files changed, 44 insertions(+), 10 deletions(-) -- 1.7.10.4

3 4

[PATCH] Add the iSCSI dependency packages to README
by Crístian Viana 20 Jan '14

20 Jan '14

The README file contains instructions to build Kimchi from the source tree, including the exact commands (e.g. yum, apt-get, zypper) to install the dependencies on the main Linux distros. But none of those commands contain the iSCSI package which is needed to validate some types of storage pool devices. Add the iSCSI package ("iscsi-initiator-utils" or "open-iscsi", depending on the distro) to the install commands in the README file. Signed-off-by: Crístian Viana <vianac(a)linux.vnet.ibm.com> --- docs/README.md | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/docs/README.md b/docs/README.md index 1ccea01..d80486b 100644 --- a/docs/README.md +++ b/docs/README.md @@ -37,7 +37,8 @@ Install Dependencies libvirt libxml2-python python-imaging \ PyPAM m2crypto python-jsonschema rpm-build \ qemu-kvm python-psutil python-ethtool sos \ - python-ipaddr python-lxml nfs-utils + python-ipaddr python-lxml nfs-utils \ + iscsi-initiator-utils # If using RHEL6, install the following additional packages: $ sudo yum install python-unittest2 python-ordereddict # Restart libvirt to allow configuration changes to take effect @@ -58,7 +59,8 @@ for more information on how to configure your system to access this repository. libvirt-bin python-libxml2 python-imaging \ python-pam python-m2crypto python-jsonschema \ qemu-kvm libtool python-psutil python-ethtool \ - sosreport python-ipaddr python-lxml nfs-common + sosreport python-ipaddr python-lxml nfs-common \ + open-iscsi Packages version requirement: python-jsonschema >= 1.3.0 @@ -71,7 +73,7 @@ for more information on how to configure your system to access this repository. libvirt python-libxml2 python-imaging \ python-pam python-M2Crypto python-jsonschema \ rpm-build kvm python-psutil python-ethtool \ - python-ipaddr python-lxml nfs-client + python-ipaddr python-lxml nfs-client open-iscsi Packages version requirement: python-psutil >= 0.6.0 -- 1.8.4.2

3 3

[RFC]: a new mechanism to improve controller
by Sheldon 20 Jan '14

20 Jan '14

as we all know we will use the follow URL for vm device. /vms/vm-name/devices/device-name we will create a new path "vm" for these devices in source tree. all device collections and resources will be in this path. |-- src | |-- distros.d | | `-- ubuntu.json | |-- kimchi | | |-- control | | | |-- base.py | | | |-- vm | | | | |-- ifaces.py | | | | `-- cdroms.py Now, I'd like to introduce a new mechanism to dispatch the vm devices(sub-collections). you can see this patch [PATCH 2/7] add a method to load vms sub collection automatically After this patch, we can dispatch the vm sub-collections more easily. For example, if we want to add a sub-collection ifaces, we can just define it as follow in iface.py +@SubCollection('ifaces') +class VmIfaces(Collection): + def __init__(self, model, vm): + super(VmIfaces, self).__init__(model) + self.resource = VmIface + self.vm = vm + self.resource_args = [self.vm, ] + self.model_args = [self.vm, ] We do not need to touch the vm.py any more. The tag @SubCollection('ifaces') means this is sub-collection of VM and the URL is /vms/vm-name/ifaces/iface-name discuss with Mark and Royce, they think we can improve the root also by this way. Such as we can define vms collection in vms.py as follow: +@SubCollection('vms', auth=True) +class VMs(Collection): + def __init__(self, model, vm): + super(Vms, self).__init__(model) + self.resource = vm For @SubCollection('vms', auth=True): the "vms" means the URL is /vms and "auth=True" means it will set tools.kimchiauth.on ad True. After we tag the Collection with the tag @SubCollection('vms', auth=True), We do not need to touch root.py and server.py any more when we add new Collections. And we can remove the follow codes. diff --git a/src/kimchi/root.py b/src/kimchi/root.py index 3cc6321..2bac4e1 100644 --- a/src/kimchi/root.py +++ b/src/kimchi/root.py @@ -27,19 +27,6 @@ import json from kimchi import auth from kimchi import template -from kimchi.config import get_api_schema_file -from kimchi.control.base import Resource -from kimchi.control.config import Config -from kimchi.control.debugreports import DebugReports -from kimchi.control.host import Host -from kimchi.control.interfaces import Interfaces -from kimchi.control.networks import Networks -from kimchi.control.plugins import Plugins -from kimchi.control.storagepools import StoragePools -from kimchi.control.tasks import Tasks -from kimchi.control.templates import Templates -from kimchi.control.utils import parse_request -from kimchi.control.vms import VMs from kimchi.exception import OperationFailed @@ -57,16 +44,6 @@ class Root(Resource): for key in self._handled_error]) Resource.__init__(self, model) - self.vms = VMs(model) - self.templates = Templates(model) - self.storagepools = StoragePools(model) - self.interfaces = Interfaces(model) - self.networks = Networks(model) - self.tasks = Tasks(model) - self.config = Config(model) - self.host = Host(model) - self.debugreports = DebugReports(model) - self.plugins = Plugins(model) self.api_schema = json.load(open(get_api_schema_file())) def error_production_handler(self, status, message, traceback, version): diff --git a/src/kimchi/server.py b/src/kimchi/server.py index b820263..08dfde2 100644 --- a/src/kimchi/server.py +++ b/src/kimchi/server.py @@ -74,13 +74,6 @@ class Server(object): 'tools.sessions.storage_type': 'file', 'tools.sessions.storage_path': config.get_session_path(), 'tools.kimchiauth.on': False}, - '/host': {'tools.kimchiauth.on': True}, - '/vms': {'tools.kimchiauth.on': True}, - '/templates': {'tools.kimchiauth.on': True}, - '/networks': {'tools.kimchiauth.on': True}, - '/storagepools': {'tools.kimchiauth.on': True}, - '/tasks': {'tools.kimchiauth.on': True}, - '/debugreports': {'tools.kimchiauth.on': True}, '/css': { 'tools.staticdir.on': True, 'tools.staticdir.dir': 'ui/css', -- Thanks and best regards! Sheldon Feng(冯少合)<shaohef(a)linux.vnet.ibm.com> IBM Linux Technology Center

2 1