[Starlingx-discuss] glance api service is not started in all-in-one deployment
Hi I deployed an starlingx all-in-one environment, and it seems glance api service has not been started. I also checked other services such as cinder api / cinder volume, there are started. /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-schedule start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-volume start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-backup start But these are not started by service file such as openstack-cinder-api.service / openstack-cinder-scheduler.service. This is really different with the deployment by devstack. Could anybody point out how the services are started? How to restart cinder/glance service in starlingx deployment? Thanks. Regards Liang
Hi, On StarlingX, glance-api is managed by SM (short for service management). You can see a list of all the services by using "sm-dump" (from a root prompt or use sudo). If it is working, the sate should be "enabled-active". To restart the service us "sm-restart service glance-api". Also check the logs (/var/log/glance/glance-api.log) for any obvious errors that prevent the service from restarting. ________________________________ From: Fang, Liang A [liang.a.fang@intel.com] Sent: Friday, November 02, 2018 11:34 AM To: starlingx-discuss@lists.starlingx.io Subject: [Starlingx-discuss] glance api service is not started in all-in-one deployment Hi I deployed an starlingx all-in-one environment, and it seems glance api service has not been started. I also checked other services such as cinder api / cinder volume, there are started. /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-schedule start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-volume start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-backup start But these are not started by service file such as openstack-cinder-api.service / openstack-cinder-scheduler.service. This is really different with the deployment by devstack. Could anybody point out how the services are started? How to restart cinder/glance service in starlingx deployment? Thanks. Regards Liang
Thanks Stefan. Is there any help doc about SM? Where does SM read the service list and the service exec from? Regards Liang From: Dinescu, Stefan [mailto:Stefan.Dinescu@windriver.com] Sent: Friday, November 2, 2018 6:11 PM To: Fang, Liang A <liang.a.fang@intel.com>; starlingx-discuss@lists.starlingx.io Subject: RE: [Starlingx-discuss] glance api service is not started in all-in-one deployment Hi, On StarlingX, glance-api is managed by SM (short for service management). You can see a list of all the services by using "sm-dump" (from a root prompt or use sudo). If it is working, the sate should be "enabled-active". To restart the service us "sm-restart service glance-api". Also check the logs (/var/log/glance/glance-api.log) for any obvious errors that prevent the service from restarting. ________________________________ From: Fang, Liang A [liang.a.fang@intel.com] Sent: Friday, November 02, 2018 11:34 AM To: starlingx-discuss@lists.starlingx.io<mailto:starlingx-discuss@lists.starlingx.io> Subject: [Starlingx-discuss] glance api service is not started in all-in-one deployment Hi I deployed an starlingx all-in-one environment, and it seems glance api service has not been started. I also checked other services such as cinder api / cinder volume, there are started. /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-schedule start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-volume start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-backup start But these are not started by service file such as openstack-cinder-api.service / openstack-cinder-scheduler.service. This is really different with the deployment by devstack. Could anybody point out how the services are started? How to restart cinder/glance service in starlingx deployment? Thanks. Regards Liang
Hi, Liang Sm has its sqlite database to have all service and related information stored, the db can be found at /var/run/sm/sm.db on a controller. A quick way to list all services is to use smc command. source /etc/nova/openrc smc service-list This command is similar to sm-dump but sm-dump is only for designers to debug/troubleshooting. smc also provides other sub commands to for querying service, servicegroup, and servicenode. Regards, Bin From: Fang, Liang A [liang.a.fang@intel.com] Sent: Saturday, November 03, 2018 12:43 AM To: Dinescu, Stefan; starlingx-discuss@lists.starlingx.io Subject: Re: [Starlingx-discuss] glance api service is not started in all-in-one deployment Thanks Stefan. Is there any help doc about SM? Where does SM read the service list and the service exec from? Regards Liang From: Dinescu, Stefan [mailto:Stefan.Dinescu@windriver.com] Sent: Friday, November 2, 2018 6:11 PM To: Fang, Liang A <liang.a.fang@intel.com>; starlingx-discuss@lists.starlingx.io Subject: RE: [Starlingx-discuss] glance api service is not started in all-in-one deployment Hi, On StarlingX, glance-api is managed by SM (short for service management). You can see a list of all the services by using "sm-dump" (from a root prompt or use sudo). If it is working, the sate should be "enabled-active". To restart the service us "sm-restart service glance-api". Also check the logs (/var/log/glance/glance-api.log) for any obvious errors that prevent the service from restarting. ________________________________ From: Fang, Liang A [liang.a.fang@intel.com] Sent: Friday, November 02, 2018 11:34 AM To: starlingx-discuss@lists.starlingx.io<mailto:starlingx-discuss@lists.starlingx.io> Subject: [Starlingx-discuss] glance api service is not started in all-in-one deployment Hi I deployed an starlingx all-in-one environment, and it seems glance api service has not been started. I also checked other services such as cinder api / cinder volume, there are started. /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-schedule start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-volume start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-backup start But these are not started by service file such as openstack-cinder-api.service / openstack-cinder-scheduler.service. This is really different with the deployment by devstack. Could anybody point out how the services are started? How to restart cinder/glance service in starlingx deployment? Thanks. Regards Liang
Thank you Bin. Regards Liang From: Qian, Bin [mailto:Bin.Qian@windriver.com] Sent: Tuesday, November 6, 2018 3:46 AM To: Fang, Liang A <liang.a.fang@intel.com>; Dinescu, Stefan <Stefan.Dinescu@windriver.com>; starlingx-discuss@lists.starlingx.io Subject: RE: [Starlingx-discuss] glance api service is not started in all-in-one deployment Hi, Liang Sm has its sqlite database to have all service and related information stored, the db can be found at /var/run/sm/sm.db on a controller. A quick way to list all services is to use smc command. source /etc/nova/openrc smc service-list This command is similar to sm-dump but sm-dump is only for designers to debug/troubleshooting. smc also provides other sub commands to for querying service, servicegroup, and servicenode. Regards, Bin From: Fang, Liang A [liang.a.fang@intel.com] Sent: Saturday, November 03, 2018 12:43 AM To: Dinescu, Stefan; starlingx-discuss@lists.starlingx.io<mailto:starlingx-discuss@lists.starlingx.io> Subject: Re: [Starlingx-discuss] glance api service is not started in all-in-one deployment Thanks Stefan. Is there any help doc about SM? Where does SM read the service list and the service exec from? Regards Liang From: Dinescu, Stefan [mailto:Stefan.Dinescu@windriver.com] Sent: Friday, November 2, 2018 6:11 PM To: Fang, Liang A <liang.a.fang@intel.com<mailto:liang.a.fang@intel.com>>; starlingx-discuss@lists.starlingx.io<mailto:starlingx-discuss@lists.starlingx.io> Subject: RE: [Starlingx-discuss] glance api service is not started in all-in-one deployment Hi, On StarlingX, glance-api is managed by SM (short for service management). You can see a list of all the services by using "sm-dump" (from a root prompt or use sudo). If it is working, the sate should be "enabled-active". To restart the service us "sm-restart service glance-api". Also check the logs (/var/log/glance/glance-api.log) for any obvious errors that prevent the service from restarting. ________________________________ From: Fang, Liang A [liang.a.fang@intel.com] Sent: Friday, November 02, 2018 11:34 AM To: starlingx-discuss@lists.starlingx.io<mailto:starlingx-discuss@lists.starlingx.io> Subject: [Starlingx-discuss] glance api service is not started in all-in-one deployment Hi I deployed an starlingx all-in-one environment, and it seems glance api service has not been started. I also checked other services such as cinder api / cinder volume, there are started. /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-schedule start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-volume start /bin/sh /usr/lib/ocf/resource.d/openstack/cinder-backup start But these are not started by service file such as openstack-cinder-api.service / openstack-cinder-scheduler.service. This is really different with the deployment by devstack. Could anybody point out how the services are started? How to restart cinder/glance service in starlingx deployment? Thanks. Regards Liang
participants (3)
-
Dinescu, Stefan
-
Fang, Liang A
-
Qian, Bin