March 2015 - Kimchi-devel - oVirt List Archives

[PATCH V2] Update host number of cpus and total physical memory
by Jose Ricardo Ziviani 01 Apr '15

01 Apr '15

- The number of cpus and the total physical memory information displayed in the host section can change, so them cannot be static. This commit updates those values whenever a GET request hits the backend looking for host information. --- src/kimchi/model/host.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/src/kimchi/model/host.py b/src/kimchi/model/host.py index 4419bb3..588bb28 100644 --- a/src/kimchi/model/host.py +++ b/src/kimchi/model/host.py @@ -87,8 +87,8 @@ class HostModel(object): res['cpu_model'] = line.split(':')[1].strip() break - res['cpus'] = psutil.NUM_CPUS - res['memory'] = psutil.TOTAL_PHYMEM + res['cpus'] = 0 + res['memory'] = 0L # Include IBM PowerKVM name to supported distro names _sup_distros = platform._supported_dists + ('ibm_powerkvm',) @@ -102,6 +102,8 @@ class HostModel(object): return res def lookup(self, *name): + self.host_info['cpus'] = psutil._psplatform.get_num_cpus() + self.host_info['memory'] = psutil.phymem_usage().total return self.host_info def swupdate(self, *name): -- 1.9.1

3 3

[PATCH V3] Number of CPUs and Physical Memory
by Jose Ricardo Ziviani 01 Apr '15

01 Apr '15

V3: psutil has changed its interface to retrieve the number of CPUs, this new V3 updates the code while supporting the old version (still needed). Jose Ricardo Ziviani (1): Update host number of cpus and total physical memory src/kimchi/model/host.py | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) -- 1.9.1

2 2

[PATCH][RFC] Kimchi and relative URLs
by Frédéric Bonnard 01 Apr '15

01 Apr '15

From: Frederic Bonnard <frediz(a)linux.vnet.ibm.com> Hi, many people and distros use a subdirectory in the configuration of the system's webserver to configure different websites. The usual way to setup a website is thus to : 1. use the distro's webserver to serve all the websites 2. have a subdirectory to copy the configuration file that makes available a website by only restarting the webserver (and removing the website is just done by removing the configuration file and restart). Concerning this point, the classical configurations can be : a) a sub web location : https://website.org/webapp b) a virtual host : http://webapp.website.org a) and b) correspond each to different configuration files that are just dropped into : - Ubuntu/Debian : /etc/{apache,nginx,..}/sites-available/{webapp_subweb,webapp_virthost}.conf - Fedora/Opensuse : /etc/{httpd,nginx,..}/conf.d/ At the moment kimchi is launched with a private instance of nginx, not the system's installed one and it would be nice to have this improved and this is covered in another thread : http://lists.ovirt.org/pipermail/kimchi-devel/2015-February/009642.html Provided that previous patch, one can use the distro's webserver to run proxy kimchi and that fullfills 1. Then for 2. one need a configuration file for either b) (provided for apache in the link above ; I have also one for nginx ) and a), which I provide in this patch. But this nginx configuration file is not enough for a sub web location configuration as kimchi relies on absolute path based on / from the http://server/. I tried to make URLs paths relative so that kimchi doesn't have to know where it has been placed on the webserver, for example https://server/kimchi/ which will probably be the URL used in distros. I'm not a web developer and I tried to modify all failing requests with a a) configuration with relative paths in mind. So I'd like comments on the way this is done, and this may need more extensive testing (paths I didn't test). Though I used that code back on a virtualhost configuration and it worked as well. Thanks for your help, F. Frederic Bonnard (1): Making urls relative docs/Makefile.am | 1 + docs/nginx.conf.subsite.ex | 37 +++++++ src/kimchi/screenshot.py | 2 +- ui/css/theme-default/template_add.css | 20 ++-- ui/css/theme-default/topbar.css | 2 +- ui/js/src/kimchi.api.js | 176 +++++++++++++++++----------------- ui/js/src/kimchi.login.js | 2 +- ui/pages/guest.html.tmpl | 2 +- ui/pages/help/dita-help.xsl | 4 +- ui/pages/kimchi-ui.html.tmpl | 4 +- ui/pages/storagepool-add.html.tmpl | 2 +- ui/pages/tabs/storage.html.tmpl | 2 +- ui/pages/template-add.html.tmpl | 2 +- 13 files changed, 147 insertions(+), 109 deletions(-) create mode 100644 docs/nginx.conf.subsite.ex -- 1.9.1

4 9

[PATCH v3] test_model: Power arch fixes
by Daniel Henrique Barboza 01 Apr '15

01 Apr '15

changes in v3: - rebase to latest master changes in v2: - changed iso_path to reflect the new ISO being created (Ubuntu 14.04) Daniel Henrique Barboza (1): test/test_model: Power architecture fixes tests/iso_gen.py | 3 ++- tests/test_model.py | 12 +++++++----- 2 files changed, 9 insertions(+), 6 deletions(-) -- 1.9.3

2 3

[kimchi-devel][PATCH] Issue #623: Fix mismatch in host device details query
by lvroyce＠linux.vnet.ibm.com 01 Apr '15

01 Apr '15

From: Royce Lv <lvroyce(a)linux.vnet.ibm.com> Commit #902d2957 began to use splitted devices names, while get detailed device still use original names. Mismatched name results in error raised when listing host devices. Fix it by using splitted device name as well. Signed-off-by: Royce Lv <lvroyce(a)linux.vnet.ibm.com> --- src/kimchi/disks.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/kimchi/disks.py b/src/kimchi/disks.py index a092c17..5cf6885 100644 --- a/src/kimchi/disks.py +++ b/src/kimchi/disks.py @@ -64,7 +64,7 @@ def _get_dev_major_min(name): dev_list = _get_lsblk_devs(keys) for dev in dev_list: - if dev['name'] == name: + if dev['name'].split()[0] == name: maj_min = dev['maj:min'] break else: -- 1.9.1

2 2

[PATCH 0/3] Fix backend VM, POOL, NETWORK naming validation
by Rodrigo Trujillo 01 Apr '15

01 Apr '15

This patchset fixes the backend JSON name checkings when user creates a new VM, POOL or NETWORK. Rodrigo Trujillo (3): Prohibits slashes '/' in VM name Move slash checking in storagepool name from UI to backend Fix Network create name checking in backend (slashes and quotes) src/kimchi/API.json | 4 ++++ src/kimchi/i18n.py | 6 +++--- src/kimchi/model/utils.py | 3 ++- ui/js/src/kimchi.storagepool_add_main.js | 5 ----- ui/pages/i18n.json.tmpl | 1 - 5 files changed, 9 insertions(+), 10 deletions(-) -- 2.1.0

2 7

[PATCH] Github issue #630: fixing None.strip() error
by Daniel Henrique Barboza 30 Mar '15

30 Mar '15

This simple patch fixes the scenario in repositories.py where baseurl or mirrorlist is None and the code was executing baseurl.strip() and mirrorlist.strip(), leading to errors in tests/test_model.py (test_repository_update) and probably other bugs. Signed-off-by: Daniel Henrique Barboza <dhbarboza82(a)gmail.com> --- src/kimchi/repositories.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/kimchi/repositories.py b/src/kimchi/repositories.py index f1e1eb3..19c5810 100644 --- a/src/kimchi/repositories.py +++ b/src/kimchi/repositories.py @@ -275,10 +275,10 @@ class YumRepo(object): config = params.get('config', {}) mirrorlist = config.get('mirrorlist', None) - if len(baseurl.strip()) == 0: + if baseurl is not None and len(baseurl.strip()) == 0: baseurl = None - if len(mirrorlist.strip()) == 0: + if mirrorlist is not None and len(mirrorlist.strip()) == 0: mirrorlist = None if baseurl is None and mirrorlist is None: -- 1.9.3

2 3

[PATCH] issue #628: Set ref_cnt=1 on cloned disks
by Crístian Viana 30 Mar '15

30 Mar '15

Cloned disks currently don't have ref_cnt stored locally, so they may be attached again to another VM - or even to the same one. When cloning a VM, set their disks' ref_cnt to 1 to prevent them from being reattached. Fix issue #628 (Storage volume ref_cnt is not updated after vm clone). Signed-off-by: Crístian Viana <vianac(a)linux.vnet.ibm.com> --- src/kimchi/model/vms.py | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/src/kimchi/model/vms.py b/src/kimchi/model/vms.py index 86c173f..af5e765 100644 --- a/src/kimchi/model/vms.py +++ b/src/kimchi/model/vms.py @@ -449,6 +449,10 @@ class VMModel(object): vir_conn = self.conn.get() + def _delete_disk_from_objstore(path): + with self.objstore as session: + session.delete('storagevolume', path) + for i, path in enumerate(all_paths): try: vir_orig_vol = vir_conn.storageVolLookupByPath(path) @@ -518,6 +522,11 @@ class VMModel(object): xml = xml_item_update(xml, XPATH_DOMAIN_DISK_BY_FILE % path, new_vol['path'], 'file') + # set the new disk's ref_cnt + with self.objstore as session: + session.store('storagevolume', new_vol['path'], {'ref_cnt': 1}) + rollback.prependDefer(_delete_disk_from_objstore, new_vol['path']) + # remove the new volume should an error occur later rollback.prependDefer(self.storagevolume.delete, new_pool_name, new_vol_name) -- 2.1.0

2 2

[PATCH 0/3] Fix issue #461 (backend)
by Crístian Viana 30 Mar '15

30 Mar '15

This patchset fixes the backend part of issue #461. The UI should also be updated in order to use the metadata property. Other related fixes have been committed as well. Crístian Viana (3): Handle empty variables when updating YUM repository Use more generic message in repo mirror list error issue #461: Add 'metalink' support for YUM repositories src/kimchi/API.json | 5 +++++ src/kimchi/i18n.py | 5 +++-- src/kimchi/repositories.py | 25 +++++++++++++++++++------ 3 files changed, 27 insertions(+), 8 deletions(-) -- 2.1.0

3 9

RFC - New option 'Serial Console' in the template UI
by Jose Ricardo Ziviani 30 Mar '15

30 Mar '15

Hello everybody! I'd like to hear from you if it would be nice to have a new option in the Template UI to redirect the output to a serial console. Today, there is a field called 'Graphics' with two options [VNC, Spice]. What if we have [VNC, Spice, Serial Console (no-graphic)]? A drawback is the need of an external tool (like virsh console) to be able to take the console, as Kimchi doesn't provide one. What do you think? Thank you! -- Jose Ricardo Ziviani ----------------------------- Software Engineer Linux Technology Center - IBM

7 9