The kubelet logs are in /var/log/daemon.log

From: Venkata Ramana Veldanda <Venkata.Veldanda@radisys.com>
Sent: Friday, March 05, 2021 4:30 AM
To: starlingx-discuss@lists.starlingx.io
Cc: Srinivas Sadagopan <Srinivas.Sadagopan@radisys.com>
Subject: Re: [Starlingx-discuss] stx 4.0.1 kubernetes sriov resources suddenly disappears

[Please note: This e-mail is from an EXTERNAL e-mail address]

Adding more info to this question as this is becoming a bit show stopper for us. Any help is appreciated.

Until SRIOV everything seems to be configured properly and they reflecting under /etc/pcidp/config.json

When we compared the faulty resource intel_fpga_fec with another stable SRIOV resource ‘pci_sriov_net_datanetmh1’ the log doesn’t show any issues nor it is different. However from kubelet perspective there seems to be something wrong and resources are not showing up in “kubectl describe nodes”. Attached the log comparison.

How do we dump the kubelet logs? The command ‘journalctl -u kubelet’ works on my CentOS Kubernetes but not on StarlingX

[root@local ~]# journalctl -u kubelet

-- No entries --

Also raised a launch pad bug with logs https://bugs.launchpad.net/starlingx/+bug/1917857

From: Venkata Ramana Veldanda
Sent: Wednesday, March 3, 2021 9:50 PM
To: starlingx-discuss@lists.starlingx.io
Cc: Srinivas Sadagopan <Srinivas.Sadagopan@radisys.com>
Subject: stx 4.0.1 kubernetes sriov resources suddenly disappears

Hi,

We are using STX 4.0.1 to install our Flexran based 5G solution in AIO-SX mode . We had created VFs on the N3000 device and some of the NIC interfaces. These resources were reflected in the Kuberenets allocatable resources. During the course of using the system, the allocatable resources for N3000 and one of the NIC interface crad started coming up as 0. The following is a part of kubectl describe nodes output. The affected resources are intel.com/intel_fpga_fec, intel.com/pci_sriov_net_datanet_c, and intel.com/pci_sriov_net_datanet_u

We already tried lock/unlock, delete and re-create the resources, but none of these help to recover the resources

[root@controller-0 sysadmin(keystone_admin)]# cat /etc/build.info

###

### StarlingX

### Release 20.06

###

OS="centos"

SW_VERSION="20.06"

BUILD_TARGET="Host Installer"

BUILD_TYPE="Formal"

BUILD_ID="r/stx.4.0"

JOB="STX_4.0_build_layer_flock"

BUILD_BY="starlingx.build@cengn.ca"

BUILD_NUMBER="22"

BUILD_HOST="starlingx_mirror"

BUILD_DATE="2020-08-05 12:25:52 +0000"

FLOCK_OS="centos"

FLOCK_JOB="STX_4.0_build_layer_flock"

FLOCK_BUILD_BY="starlingx.build@cengn.ca"

FLOCK_BUILD_NUMBER="22"

FLOCK_BUILD_HOST="starlingx_mirror"

FLOCK_BUILD_DATE="2020-08-05 12:25:52 +0000"

Capacity:

cpu: 96

ephemeral-storage: 10190100Ki

hugepages-1Gi: 46Gi

hugepages-2Mi: 0

intel.com/intel_fpga_fec: 0

intel.com/pci_sriov_net_datanet_c: 0

intel.com/pci_sriov_net_datanet_u: 0

intel.com/pci_sriov_net_datanetbh1: 8

intel.com/pci_sriov_net_datanetdn1: 8

intel.com/pci_sriov_net_datanetmh1: 8

memory: 97436728Ki

pods: 110

Allocatable:

cpu: 92

ephemeral-storage: 9391196145

hugepages-1Gi: 46Gi

hugepages-2Mi: 0

intel.com/intel_fpga_fec: 0

intel.com/pci_sriov_net_datanet_c: 0

intel.com/pci_sriov_net_datanet_u: 0

intel.com/pci_sriov_net_datanetbh1: 8

intel.com/pci_sriov_net_datanetdn1: 8

intel.com/pci_sriov_net_datanetmh1: 8

It seems like everything is "OK" upto sriov device plugin because the sriov device plugin pod logs do show that the correct number of resources are getting updated towards kubernetes

Here are some of the CLI outputs:

[root@controller-0 sysadmin(keystone_admin)]# system host-device-show controller-0 pci_0000_1d_00_0

+-----------------------+---------------------------------------------------------------------------------------------------------+

| Property | Value |

+-----------------------+---------------------------------------------------------------------------------------------------------+

| name | pci_0000_1d_00_0 |

| address | 0000:1d:00.0 |

| class id | 120000 |

| vendor id | 8086 |

| device id | 0d8f |

| class name | Processing accelerators |

| vendor name | Intel Corporation |

| device name | Device 0d8f |

| numa_node | 0 |

| enabled | True |

| sriov_totalvfs | 8 |

| sriov_numvfs | 8 |

| sriov_vfs_pci_address | 0000:1d:00.1,0000:1d:00.2,0000:1d:00.3,0000:1d:00.4,0000:1d:00.5,0000:1d:00.6,0000:1d:00.7,0000:1d:01.0 |

| sriov_vf_pdevice_id | 0d90 |

| extra_info | |

| created_at | 2021-03-03T13:46:26.363470+00:00 |

| updated_at | 2021-03-03T13:47:12.684827+00:00 |

| root_key | None |

| revoked_key_ids | None |

| boot_page | None |

| bitstream_id | None |

| bmc_build_version | None |

| bmc_fw_version | None |

| driver | igb_uio |

| sriov_vf_driver | igb_uio |

+-----------------------+---------------------------------------------------------------------------------------------------------+

[root@controller-0 sysadmin(keystone_admin)]# system host-if-show controller-0 sriovfh1

+-----------------+--------------------------------------+

| Property | Value |

+-----------------+--------------------------------------+

| ifname | sriovfh1 |

| iftype | ethernet |

| ports | [u'enp177s0f3'] |

| imac | 40:a6:b7:34:e4:a3 |

| imtu | 9216 |

| ifclass | pci-sriov |

| ptp_role | none |

| aemode | None |

| schedpolicy | None |

| txhashpolicy | None |

| uuid | 6f30a690-2414-424f-b5fc-d324d63cc502 |

| ihost_uuid | 8075e0db-4cc5-4d74-8601-849adce97b7e |

| vlan_id | None |

| uses | [] |

| used_by | [] |

| created_at | |

| updated_at | |

| sriov_numvfs | 16 |

| sriov_vf_driver | vfio |

| accelerated | [True] |

+-----------------+--------------------------------------+

[root@controller-0 sysadmin(keystone_admin)]# system interface-datanetwork-list controller-0

+--------------+--------------------------------------+----------+------------------+

+--------------+--------------------------------------+----------+------------------+

+--------------+--------------------------------------+----------+------------------+

[root@controller-0 sysadmin(keystone_admin)]# system interface-datanetwork-show datanet-c controller-0 datanet-ccontroller-0 datanet-c datanet-clist controller-0

+--------------+--------------------------------------+----------+------------------+

+--------------+--------------------------------------+----------+------------------+

+--------------+--------------------------------------+----------+------------------+

[root@controller-0 sysadmin(keystone_admin)]# system interface-datanetwork-list controller-0show 63a44e7b-18f4-4f9b-8504-a950cb8abb86

+------------------+--------------------------------------+

| Property | Value |

+------------------+--------------------------------------+

| hostname | controller-0 |

| uuid | 63a44e7b-18f4-4f9b-8504-a950cb8abb86 |

| ifname | sriovfh1 |

| datanetwork_name | datanet-c |

+------------------+--------------------------------------+

[root@controller-0 sysadmin(keystone_admin)]# system interface-datanetwork-show 63a44e7b-18f4-4f9b-8504-a950cb8abb86e155e1d0-8dec-47e6-ac60-076832698a95

+------------------+--------------------------------------+

| Property | Value |

+------------------+--------------------------------------+

| hostname | controller-0 |

| uuid | e155e1d0-8dec-47e6-ac60-076832698a95 |

| ifname | sriovfh1 |

| datanetwork_name | datanet-u |

+------------------+--------------------------------------+

[root@controller-0 sysadmin(keystone_admin)]#

sriov device plugin logs:

=====================================================================================

controller-0:/home/sysadmin# kubectl logs kube-sriov-device-plugin-amd64-h9kp8 -n kube-system | grep intel_fpga_fec

"resourceName": "intel_fpga_fec",

I0303 13:47:07.275600 138581 manager.go:106] unmarshalled ResourceList: [{ResourcePrefix: ResourceName:pci_sriov_net_datanet_c DeviceType:netDevice Selectors:0xc00000c900 SelectorObj:0xc0002601e0} {ResourcePrefix: ResourceName:pci_sriov_net_datanetbh1 DeviceType:netDevice Selectors:0xc00000c920 SelectorObj:0xc000260280} {ResourcePrefix: ResourceName:pci_sriov_net_datanetdn1 DeviceType:netDevice Selectors:0xc00000c940 SelectorObj:0xc000260320} {ResourcePrefix: ResourceName:pci_sriov_net_datanetmh1 DeviceType:netDevice Selectors:0xc00000c960 SelectorObj:0xc0002603c0} {ResourcePrefix: ResourceName:pci_sriov_net_datanet_u DeviceType:netDevice Selectors:0xc00000c9a0 SelectorObj:0xc000260460} {ResourcePrefix: ResourceName:intel_fpga_fec DeviceType:accelerator Selectors:0xc00000c9c0 SelectorObj:0xc00014c820}]

I0303 13:47:07.275709 138581 manager.go:193] validating resource name "intel.com/intel_fpga_fec"

I0303 13:47:07.450400 138581 manager.go:116] Creating new ResourcePool: intel_fpga_fec

I0303 13:47:07.450446 138581 manager.go:145] New resource server is created for intel_fpga_fec ResourcePool

I0303 13:47:07.453772 138581 server.go:191] starting intel_fpga_fec device plugin endpoint at: intel.com_intel_fpga_fec.sock

I0303 13:47:07.454032 138581 server.go:217] intel_fpga_fec device plugin endpoint started serving

I0303 13:47:07.640208 138581 server.go:106] Plugin: intel.com_intel_fpga_fec.sock gets registered successfully at Kubelet

I0303 13:47:07.640225 138581 server.go:131] ListAndWatch(intel_fpga_fec) invoked

I0303 13:47:07.640342 138581 server.go:139] ListAndWatch(intel_fpga_fec): send devices &ListAndWatchResponse{Devices:[]*Device{&Device{ID:0000:1d:01.0,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:0,},},},},&Device{ID:0000:1d:00.1,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:0,},},},},&Device{ID:0000:1d:00.2,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:0,},},},},&Device{ID:0000:1d:00.3,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:0,},},},},&Device{ID:0000:1d:00.4,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:0,},},},},&Device{ID:0000:1d:00.5,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:0,},},},},&Device{ID:0000:1d:00.6,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:0,},},},},&Device{ID:0000:1d:00.7,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:0,},},},},},}

controller-0:/home/sysadmin#

controller-0:/home/sysadmin# kubectl logs kube-sriov-device-plugin-amd64-h9kp8 -n kube-system | grep intel_fpga_fecpci_sriov_net_datanet_c

"resourceName": "pci_sriov_net_datanet_c",

I0303 13:47:07.275645 138581 manager.go:193] validating resource name "intel.com/pci_sriov_net_datanet_c"

I0303 13:47:07.449979 138581 manager.go:116] Creating new ResourcePool: pci_sriov_net_datanet_c

I0303 13:47:07.450122 138581 manager.go:145] New resource server is created for pci_sriov_net_datanet_c ResourcePool

I0303 13:47:07.450478 138581 server.go:191] starting pci_sriov_net_datanet_c device plugin endpoint at: intel.com_pci_sriov_net_datanet_c.sock

I0303 13:47:07.451088 138581 server.go:217] pci_sriov_net_datanet_c device plugin endpoint started serving

I0303 13:47:07.639929 138581 server.go:131] ListAndWatch(pci_sriov_net_datanet_c) invoked

I0303 13:47:07.640068 138581 server.go:106] Plugin: intel.com_pci_sriov_net_datanet_c.sock gets registered successfully at Kubelet

I0303 13:47:07.639996 138581 server.go:139] ListAndWatch(pci_sriov_net_datanet_c): send devices &ListAndWatchResponse{Devices:[]*Device{&Device{ID:0000:b1:0f.4,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.2,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.4,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.0,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.1,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.3,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.6,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.1,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.3,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.5,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.6,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.2,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.5,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.7,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.7,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.0,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},},}

controller-0:/home/sysadmin# kubectl logs kube-sriov-device-plugin-amd64-h9kp8 -n kube-system | grep pci_sriov_net_datanet_cu

"resourceName": "pci_sriov_net_datanet_u",

I0303 13:47:07.275700 138581 manager.go:193] validating resource name "intel.com/pci_sriov_net_datanet_u"

I0303 13:47:07.450306 138581 manager.go:116] Creating new ResourcePool: pci_sriov_net_datanet_u

I0303 13:47:07.450388 138581 manager.go:145] New resource server is created for pci_sriov_net_datanet_u ResourcePool

I0303 13:47:07.453443 138581 server.go:191] starting pci_sriov_net_datanet_u device plugin endpoint at: intel.com_pci_sriov_net_datanet_u.sock

I0303 13:47:07.453750 138581 server.go:217] pci_sriov_net_datanet_u device plugin endpoint started serving

I0303 13:47:07.639929 138581 server.go:131] ListAndWatch(pci_sriov_net_datanet_u) invoked

I0303 13:47:07.639945 138581 server.go:139] ListAndWatch(pci_sriov_net_datanet_u): send devices &ListAndWatchResponse{Devices:[]*Device{&Device{ID:0000:b1:0e.1,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.5,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.2,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.3,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.0,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.0,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.1,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.3,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.5,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.6,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.7,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.4,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.6,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0e.7,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.2,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},&Device{ID:0000:b1:0f.4,Health:Healthy,Topology:&TopologyInfo{Nodes:[]*NUMANode{&NUMANode{ID:1,},},},},},}

I0303 13:47:07.640791 138581 server.go:106] Plugin: intel.com_pci_sriov_net_datanet_u.sock gets registered successfully at Kubelet

controller-0:/home/sysadmin#

Venkata Veldanda