[Starlingx-discuss] About execute bash command in sysinv conductor

Wensley, Barton Barton.Wensley at windriver.com
Tue Jan 14 13:57:36 UTC 2020


Thanks Tingjie. Using kubectl for prototyping is OK, but when you are submitting code for review, please extend the KubeOperator to support the APIs you require.

Bart

From: Chen, Tingjie [mailto:tingjie.chen at intel.com]
Sent: January 14, 2020 8:27 AM
To: Wensley, Barton; starlingx-discuss at lists.starlingx.io
Subject: RE: About execute bash command in sysinv conductor

Thanks Barton,

I think it is the reason for the failure :)

Yes, It is a better way to use Kubernetes.client API to communicate with Kubernetes cluster, but the interface in sysinv/common/Kubernetes.py is not enough to support the Ceph plugin, so I have made a simple kube Api to verify the functionality and it will be better to extend the interface in Kubernetes.py to support Ceph plugin further, for example: https://github.com/kubernetes-client/python/blob/master/kubernetes/docs/CoreV1Api.md#connect_get_namespaced_pod_exec

Thanks,
Tingjie

From: Wensley, Barton <Barton.Wensley at windriver.com>
Sent: Tuesday, January 14, 2020 8:58 PM
To: Chen, Tingjie <tingjie.chen at intel.com>; starlingx-discuss at lists.starlingx.io
Subject: RE: About execute bash command in sysinv conductor

Tingjie,

It looks like you are missing the --kubeconfig parameter for kubectl, which should point to the /etc/kubernetes/admin.conf file which contains the authentication credentials.

However, it is better to use the python-kubernetes API instead of kubectl - take a look at the KubeOperator class in sysinv/common/kubernetes.py - that is our preferred way to communicate with the kubernetes API.

Bart

From: Chen, Tingjie [mailto:tingjie.chen at intel.com]
Sent: January 14, 2020 5:47 AM
To: starlingx-discuss at lists.starlingx.io<mailto:starlingx-discuss at lists.starlingx.io>
Subject: [Starlingx-discuss] About execute bash command in sysinv conductor

Hi,

I have met an issue when add a new plugin which called by sysinv,
The plugin is to execute some bash commands, such as kubectl to get information, but it seems I have tried by using:
Subprocess.Popen() and subprocess.check_output() it return failed with non-zero exit status 1.
But I tried on my local Ubuntu environment, the result is OK.

Is there any audit or privilege or some audit in Sysinv Service?

Piece of code:
        execute_process = subprocess.Popen(command,
            stdout=subprocess.PIPE, stderr=sys.stderr)
        objects = yaml.safe_load(execute_process.stdout)
        if execute_process.wait() != 0:
            raise ApiError
        return objects

Following is the output:
-----------------------------------------------------
sysinv 2020-01-14 10:31:20.038 1172858 INFO sysinv.common.ceph [-] ceph osd crush rule ls
sysinv 2020-01-14 10:31:20.038 1172858 ERROR rook_client_kubeapi [-] ['/usr/bin/kubectl', 'get', '--namespace', 'kube-system', 'pod', '-l', 'app=rook-ceph-tools', '-o', 'jsonpath="{.items[0].metadata.name}"']
sysinv 2020-01-14 10:31:20.038 1172858 INFO rook_client_kubeapi [-] Get pod command: ['/usr/bin/kubectl', 'get', '--namespace', 'kube-system', 'pod', '-l', 'app=rook-ceph-tools', '-o', 'jsonpath="{.items[0].metadata.name}"']
sysinv 2020-01-14 10:31:45.026 1172858 ERROR sysinv.common.ceph [-] Exception in ls crush rule: Command '['/usr/bin/kubectl', 'get', '--namespace', 'kube-system', 'pod', '-l', 'app=rook-ceph-tools', '-o', 'jsonpath="{.items[0].metadata.name}"']' returned non-zero exit status 1: CalledProcessError: Command '['/usr/bin/kubectl', 'get', '--namespace', 'kube-system', 'pod', '-l', 'app=rook-ceph-tools', '-o', 'jsonpath="{.items[0].metadata.name}"']' returned non-zero exit status 1
sysinv 2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp [-] Exception during message handling: CephFailure: OSD crush rule ls failed. Command '['/usr/bin/kubectl', 'get', '--namespace', 'kube-system', 'pod', '-l', 'app=rook-ceph-tools', '-o', 'jsonpath="{.items[0].metadata.name}"']' returned non-zero exit status 1
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp Traceback (most recent call last):
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp   File "/usr/lib64/python2.7/site-packages/sysinv/openstack/common/rpc/amqp.py", line 437, in _process_data
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp     **args)
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp   File "/usr/lib64/python2.7/site-packages/sysinv/openstack/common/rpc/dispatcher.py", line 172, in dispatch
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp     result = getattr(proxyobj, method)(ctxt, **kwargs)
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp   File "/usr/lib64/python2.7/site-packages/sysinv/conductor/manager.py", line 4405, in iplatform_update_by_ihost
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp     cceph.fix_crushmap(self.dbapi)
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp   File "/usr/lib64/python2.7/site-packages/sysinv/common/ceph.py", line 947, in fix_crushmap
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp     rule_is_present, __, __ = _operator._crush_rule_status(default_ceph_tier_name)
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp   File "/usr/lib64/python2.7/site-packages/sysinv/common/ceph.py", line 67, in _crush_rule_status
2020-01-14 10:31:45.027 1172858 ERROR sysinv.openstack.common.rpc.amqp     raise exception.CephFailure("OSD crush rule ls failed. %s" %str(e))

Thanks,
Tingjie

SSP NST Storage
Tel: +86(21)88216699
Mobile: 15901876439

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.starlingx.io/pipermail/starlingx-discuss/attachments/20200114/4917ee3a/attachment-0001.html>


More information about the Starlingx-discuss mailing list