排除常见数据层(CDL)故障
1. 简介
本文将介绍在SMF环境中对通用数据层(CDL)进行故障排除的基础知识。您可以在此链接上找到的文档。
2.概述
思科通用数据层(CDL)是用于所有云本地应用的高性能下一代KV(Key-value)数据存储层。
CDL当前用作具有HA（高可用性）和Geo HA功能的状态管理组件。
CDL提供：
- 跨不同网络功能(NF)的通用数据存储层。
- 低延迟读写（在内存会话存储中）
- 当报告对同一会话的DoS（拒绝服务）攻击时，通知NF阻止用户。
- 高可用性 — 至少具有2个副本的本地冗余。
- 具有2个站点的地理冗余。
- 没有主要/次要概念所有插槽均可用于写操作。由于不进行主选举，从而缩短故障切换时间。
3.组成部分
- 终端：(cdl-ep-session-c1-d0-7c79c87d65-xpm5v)
- CDL终端是Kubernetes(K8s)POD。它被部署用于向NF客户端显示gRPC over HTTP2接口，用于处理数据库服务请求，并充当北向应用程序的入口点。
- 插槽：(cdl-slot-session-c1-m1-0)
- CDL终端支持多个插槽微服务。这些微服务是部署的K8s POD，用于向思科数据存储显示内部gRPC接口
- 每个插槽POD具有有限数量的会话。这些会话是字节数组格式的实际会话数据
- 索引：(cdl-index-session-c1-m1-0)
- 索引微服务保存索引相关的数据
- 然后，此索引数据用于从插槽微服务检索实际会话数据
- ETCD:(etcd-smf-etcd-cluster-0)
- CDL使用ETCD（开源密钥值存储）作为数据库服务发现。当Cisco Data Store EP启动、关闭或关闭时，会导致发布状态添加事件。因此，会向订用这些事件的每个POD发送通知。此外，当添加或删除关键事件时，它会刷新本地映射。
- 卡夫卡：（卡夫卡–0）
- Kafka POD在本地副本之间和跨站点复制数据以编制索引。对于跨站点复制，Kafak使用MirrorMaker。
- Mirror Maker:(mirror-maker-0)
- Mirror Maker POD将索引数据异地复制到远程CDL站点。它从远程站点获取数据并将其发布到本地Kafka站点，以便获取相应的索引实例。
示例：
master-1:~$ kubectl get pods -n smf-smf -o wide NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES cdl-ep-session-c1-d0-7889db4d87-5mln5 1/1 Running 0 80d 192.168.16.247 smf-data-worker-5 <none> <none> cdl-ep-session-c1-d0-7889db4d87-8q7hg 1/1 Running 0 80d 192.168.18.108 smf-data-worker-1 <none> <none> cdl-ep-session-c1-d0-7889db4d87-fj2nf 1/1 Running 0 80d 192.168.24.206 smf-data-worker-3 <none> <none> cdl-ep-session-c1-d0-7889db4d87-z6c2z 1/1 Running 0 34d 192.168.4.164 smf-data-worker-2 <none> <none> cdl-ep-session-c1-d0-7889db4d87-z7c89 1/1 Running 0 80d 192.168.7.161 smf-data-worker-4 <none> <none> cdl-index-session-c1-m1-0 1/1 Running 0 80d 192.168.7.172 smf-data-worker-4 <none> <none> cdl-index-session-c1-m1-1 1/1 Running 0 80d 192.168.24.241 smf-data-worker-3 <none> <none> cdl-index-session-c1-m2-0 1/1 Running 0 49d 192.168.18.116 smf-data-worker-1 <none> <none> cdl-index-session-c1-m2-1 1/1 Running 0 80d 192.168.7.173 smf-data-worker-4 <none> <none> cdl-index-session-c1-m3-0 1/1 Running 0 80d 192.168.24.197 smf-data-worker-3 <none> <none> cdl-index-session-c1-m3-1 1/1 Running 0 80d 192.168.18.107 smf-data-worker-1 <none> <none> cdl-index-session-c1-m4-0 1/1 Running 0 80d 192.168.7.158 smf-data-worker-4 <none> <none> cdl-index-session-c1-m4-1 1/1 Running 0 49d 192.168.16.251 smf-data-worker-5 <none> <none> cdl-slot-session-c1-m1-0 1/1 Running 0 80d 192.168.18.117 smf-data-worker-1 <none> <none> cdl-slot-session-c1-m1-1 1/1 Running 0 80d 192.168.24.201 smf-data-worker-3 <none> <none> cdl-slot-session-c1-m2-0 1/1 Running 0 80d 192.168.16.245 smf-data-worker-5 <none> <none> cdl-slot-session-c1-m2-1 1/1 Running 0 80d 192.168.18.123 smf-data-worker-1 <none> <none> cdl-slot-session-c1-m3-0 1/1 Running 0 34d 192.168.4.156 smf-data-worker-2 <none> <none> cdl-slot-session-c1-m3-1 1/1 Running 0 80d 192.168.18.78 smf-data-worker-1 <none> <none> cdl-slot-session-c1-m4-0 1/1 Running 0 34d 192.168.4.170 smf-data-worker-2 <none> <none> cdl-slot-session-c1-m4-1 1/1 Running 0 80d 192.168.7.177 smf-data-worker-4 <none> <none> cdl-slot-session-c1-m5-0 1/1 Running 0 80d 192.168.24.246 smf-data-worker-3 <none> <none> cdl-slot-session-c1-m5-1 1/1 Running 0 34d 192.168.4.163 smf-data-worker-2 <none> <none> cdl-slot-session-c1-m6-0 1/1 Running 0 80d 192.168.18.119 smf-data-worker-1 <none> <none> cdl-slot-session-c1-m6-1 1/1 Running 0 80d 192.168.16.228 smf-data-worker-5 <none> <none> cdl-slot-session-c1-m7-0 1/1 Running 0 80d 192.168.16.215 smf-data-worker-5 <none> <none> cdl-slot-session-c1-m7-1 1/1 Running 0 49d 192.168.4.167 smf-data-worker-2 <none> <none> cdl-slot-session-c1-m8-0 1/1 Running 0 49d 192.168.24.213 smf-data-worker-3 <none> <none> cdl-slot-session-c1-m8-1 1/1 Running 0 80d 192.168.16.253 smf-data-worker-5 <none> <none> etcd-smf-smf-etcd-cluster-0 2/2 Running 0 80d 192.168.11.176 smf-data-master-1 <none> <none> etcd-smf-smf-etcd-cluster-1 2/2 Running 0 48d 192.168.7.59 smf-data-master-2 <none> <none> etcd-smf-smf-etcd-cluster-2 2/2 Running 1 34d 192.168.11.66 smf-data-master-3 <none> <none> georeplication-pod-0 1/1 Running 0 80d 10.10.1.22 smf-data-master-1 <none> <none> georeplication-pod-1 1/1 Running 0 48d 10.10.1.23 smf-data-master-2 <none> <none> grafana-dashboard-cdl-smf-smf-77bd69cff7-qbvmv 1/1 Running 0 34d 192.168.7.41 smf-data-master-2 <none> <none> kafka-0 2/2 Running 0 80d 192.168.24.245 smf-data-worker-3 <none> <none> kafka-1 2/2 Running 0 49d 192.168.16.200 smf-data-worker-5 <none> <none> mirror-maker-0 1/1 Running 1 80d 192.168.18.74 smf-data-worker-1 <none> <none> zookeeper-0 1/1 Running 0 34d 192.168.11.73 smf-data-master-3 <none> <none> zookeeper-1 1/1 Running 0 48d 192.168.7.47 smf-data-master-2 <none> <none> zookeeper-2
CDL架构
4.配置演练
smf# show running-config cdl cdl system-id 1 /// unique across the site, system-id 1 is the primary site ID for sliceNames SMF1 SMF2 in HA GR CDL deploy cdl node-type db-data /// node label to configure the node affinity cdl enable-geo-replication true /// CDL GR Deployment with 2 RACKS cdl remote-site 2 db-endpoint host x.x.x.x /// Remote site cdl-ep configuration on site-1 db-endpoint port 8882 kafka-server x.x.x.x 10061 /// Remote site kafka configuration on site-1 exit kafka-server x.x.x.x 10061 exit exit cdl label-config session /// Configures the list of label for CDL pods endpoint key smi.cisco.com/node-type-3 endpoint value session slot map 1 key smi.cisco.com/node-type-3 value session exit slot map 2 key smi.cisco.com/node-type-3 value session exit slot map 3 key smi.cisco.com/node-type-3 value session exit slot map 4 key smi.cisco.com/node-type-3 value session exit slot map 5 key smi.cisco.com/node-type-3 value session exit slot map 6 key smi.cisco.com/node-type-3 value session exit slot map 7 key smi.cisco.com/node-type-3 value session exit slot map 8 key smi.cisco.com/node-type-3 value session exit index map 1 key smi.cisco.com/node-type-3 value session exit index map 2 key smi.cisco.com/node-type-3 value session exit index map 3 key smi.cisco.com/node-type-3 value session exit index map 4 key smi.cisco.com/node-type-3 value session exit exit cdl datastore session /// unique with in the site label-config session geo-remote-site [ 2 ] slice-names [ SMF1 SMF2 ] endpoint cpu-request 2000 endpoint go-max-procs 16 endpoint replica 5 /// number of cdl-ep pods endpoint external-ip x.x.x.x endpoint external-port 8882 index cpu-request 2000 index go-max-procs 8 index replica 2 /// number of replicas per mop for cdl-index, can not be changed after CDL deployement.
NOTE: If you need to change number of index replica, set the system mode to shutdown from respective ops-center CLI, change the replica and set the system mode to running index map 4 /// number of mops for cdl-index index write-factor 1 /// number of copies to be written before a successful response slot cpu-request 2000 slot go-max-procs 8 slot replica 2 /// number of replicas per mop for cdl-slot slot map 8 /// number of mops for cdl-slot slot write-factor 1 slot metrics report-idle-session-type true features instance-aware-notification enable true /// This enables GR failover notification features instance-aware-notification system-id 1 slice-names [ SMF1 ] exit features instance-aware-notification system-id 2 slice-names [ SMF2 ] exit exit cdl kafka replica 2 cdl kafka label-config key smi.cisco.com/node-type-3 cdl kafka label-config value session cdl kafka external-ip x.x.x.x 10061 exit cdl kafka external-ip x.x.x.x 10061 exit
5.故障排除
5.1 Pod故障
CDL的操作是简单明了的Key > Value db。
- 所有请求都进入cdl-endpoint pod。
- 在cdl-index pods中，我们存储密钥，轮询。
- 在cdl-slot中，我们存储值（会话信息），轮询。
- 我们为每个Pod映射（类型）定义备份（副本数量）。
- Kafka pod被用作运输公车。
- 镜像生成器用作到不同机架的传输总线（Geo冗余）。
每个的故障都可以转换为，也就是说，如果此类型/映射的所有Pod同时关闭：
- cdl-endpoint — 与CDL通信时出错
- cdl-index — 丢失会话数据的密钥
- cdl-slot — 丢失会话数据
- Kafka — 丢失Pod类型映射之间的同步选项
- 镜像生成器 — 丢失与其他地理位置还原和节点的同步
我们始终可以从相关Pod收集日志，因为cdl Pod日志的滚动更新速度不会太快，因此收集这些日志具有额外的价值。
Remamber tac-debug在注销时及时收集快照，并打印出所有数据，因为数据是存储的。
描述Pod
kubectl describe pod cdl-ep-session-c1-d0-7889db4d87-5mln5 -n smf-rcdn
收集Pod日志
kubectl logs cdl-ep-session-c1-d0-7c79c87d65-xpm5v -n smf-rcdn
5.2 CDL如何从会话密钥获取会话信息
在CDL内部，每个会话都有一个名为unique-keys的字段，用于标识此会话。
如果比较show subscriber supi和cdl show sessions summary slice-name slice1 db-name session filter的会话打印输出
- 结合supi的ipv4会话地址= "1#/#imsi-123969789012404:10.0.0.3"
- ddn + ip4地址= "1#/#lab:10.0.0.3"
- 结合supi的ipv6会话地址= "1#/#imsi-123969789012404:2001:db0:0:2::"
- ddn + ipv6 address from session = "1#/#lab:2001:db0:0:2::"
- smfTeid还使用N4会话密钥= "1#/#293601283"在对UPF上的错误进行故障排除时，这非常有用，您可以搜索会话日志并查找与会话相关的信息。
- supi + ebi = "1#/#imsi-123969789012404:ebi-5"
- supi + ddn= "1#/#imsi-123969789012404:lab"
[smf/data] smf# cdl show sessions summary slice-name slice1 db-name session filter { condition match key 1#/#293601283 }
Sun Mar 19 20:17:41.914 UTC+00:00
message params: {session-summary cli session {0 100 1#/#293601283 0 [{0 1#/#293601283}] [] 0 0 false 4096 [] [] 0} slice1}
session {
primary-key 1#/#imsi-123969789012404:1
unique-keys [ "1#/#imsi-123969789012404:10.0.0.3" "1#/#lab:10.0.0.3" "1#/#imsi-123969789012404:2001:db0:0:2::" "1#/#lab:2001:db0:0:2::" "1#/#293601283" "1#/#imsi-123969789012404:ebi-5" "1#/#imsi-123969789012404:lab" ]
non-unique-keys [ "1#/#roaming-status:visitor-lbo" "1#/#ue-type:nr-capable" "1#/#supi:imsi-123969789012404" "1#/#gpsi:msisdn-22331010101010" "1#/#pei:imei-123456789012381" "1#/#psid:1" "1#/#snssai:001000003" "1#/#dnn:lab" "1#/#emergency:false" "1#/#rat:nr" "1#/#access:3gpp" access "1#/#connectivity:5g" "1#/#udm-uecm:10.10.10.215" "1#/#udm-sdm:10.10.10.215" "1#/#auth-status:unauthenticated" "1#/#pcfGroupId:PCF-dnn=lab;" "1#/#policy:2" "1#/#pcf:10.10.10.216" "1#/#upf:10.10.10.150" "1#/#upfEpKey:10.10.10.150:10.10.10.202" "1#/#ipv4-addr:pool1/10.0.0.3" "1#/#ipv4-pool:pool1" "1#/#ipv4-range:pool1/10.0.0.1" "1#/#ipv4-startrange:pool1/10.0.0.1" "1#/#ipv6-pfx:pool1/2001:db0:0:2::" "1#/#ipv6-pool:pool1" "1#/#ipv6-range:pool1/2001:db0::" "1#/#ipv6-startrange:pool1/2001:db0::" "1#/#id-index:1:0:32768" "1#/#id-value:2/3" "1#/#chfGroupId:CHF-dnn=lab;" "1#/#chf:10.10.10.218" "1#/#amf:10.10.10.217" "1#/#peerGtpuEpKey:10.10.10.150:20.0.0.1" "1#/#namespace:smf" ]
flags [ flag3:peerGtpuEpKey:10.10.10.150:20.0.0.1 session-state-flag:smf_active ]
map-id 2
instance-id 1
app-instance-id 1
version 1
create-time 2023-03-19 20:14:14.381940117 +0000 UTC
last-updated-time 2023-03-19 20:14:14.943366502 +0000 UTC
purge-on-eval false
next-eval-time 2023-03-26 20:14:14 +0000 UTC
session-types [ rat_type:NR wps:non_wps emergency_call:false pdu_type:ipv4v6 dnn:lab qos_5qi_1_rat_type:NR ssc_mode:ssc_mode_1 always_on:disable fourg_only_ue:false up_state:active qos_5qi_5_rat_type:NR dcnr:disable smf_roaming_status:visitor-lbo dnn:lab:rat_type:NR ]
data-size 2866
}
[smf/data] smf#
如果将其与SMF的打印输出进行比较：
[smf/data] smf# show subscriber supi imsi-123969789012404 gr-instance 1 namespace smf
Sun Mar 19 20:25:47.816 UTC+00:00
subscriber-details
{
"subResponses": [
[
"roaming-status:visitor-lbo",
"ue-type:nr-capable",
"supi:imsi-123969789012404",
"gpsi:msisdn-22331010101010",
"pei:imei-123456789012381",
"psid:1",
"snssai:001000003",
"dnn:lab",
"emergency:false",
"rat:nr",
"access:3gpp access",
"connectivity:5g",
"udm-uecm:10.10.10.215",
"udm-sdm:10.10.10.215",
"auth-status:unauthenticated",
"pcfGroupId:PCF-dnn=lab;",
"policy:2",
"pcf:10.10.10.216",
"upf:10.10.10.150",
"upfEpKey:10.10.10.150:10.10.10.202",
"ipv4-addr:pool1/10.0.0.3",
"ipv4-pool:pool1",
"ipv4-range:pool1/10.0.0.1",
"ipv4-startrange:pool1/10.0.0.1",
"ipv6-pfx:pool1/2001:db0:0:2::",
"ipv6-pool:pool1",
"ipv6-range:pool1/2001:db0::",
"ipv6-startrange:pool1/2001:db0::",
"id-index:1:0:32768",
"id-value:2/3",
"chfGroupId:CHF-dnn=lab;",
"chf:10.10.10.218",
"amf:10.10.10.217",
"peerGtpuEpKey:10.10.10.150:20.0.0.1",
"namespace:smf",
"nf-service:smf"
]
]
}
检查SMF上的CDL状态：
cdl show status
cdl show sessions summary slice-name <slice name> | more
5.3 CDL Pod未启用
如何识别
检查describe pod输出（容器/成员/状态/原因、事件）。
kubectl describe pods -n <namespace> <failed pod name>
如何修复
- Pod处于挂起状态检查是否有任何标签值等于cdl/node-type复制副本数的k8s节点小于或等于标签值等于cdl/node-type值的k8s节点数
kubectl get nodes -l smi.cisco.com/node-type=<value of cdl/node-type, default value is 'session' in multi node setup)
- Pod处于CrashLoopBackOff故障状态检查etcd pod状态。如果etcd pod未运行，请修复etcd问题。
kubectl describe pods -n <namespace> <etcd pod name>
- Pod处于ImagePullBack故障状态。检查是否可以访问helm存储库和映像注册表。检查是否配置了所需的代理和dns服务器。
5.4 Mirror Maker面板处于初始化状态
检查描述Pod输出和Pod日志
kubectl describe pods -n <namespace> <failed pod name> kubectl logs -n <namespace> <failed pod name> [-c <container name>]
如何修复
- 检查为Kafka配置的外部IP是否正确
- 通过外部IP检查远程站点kafka的可用性
5.5 CDL索引复制不正确
如何识别
从一个站点添加的数据无法从其他站点访问。
如何修复
- 检查本地系统id配置和远程站点配置。
- 检查CDL终端和kafka在每个站点之间的可达性。
- 检查每个站点上的映射、索引副本和插槽。它可以在所有站点上完全相同。
5.6 CDL操作失败，但连接成功
如何修复
- 检查所有Pod是否处于就绪和运行状态。
- 只有当索引Pod与对等副本（本地或远程，如果可用）同步完成时，它们才处于就绪状态
- 仅当与对等副本（本地或远程，如果可用）同步完成时，插槽Pod才会处于就绪状态
- 如果至少一个插槽和一个索引Pod不可用，则终端处于NOT in ready状态。即使尚未就绪，也会从客户端接受grpc连接。
5.7从CDL提前或推迟清除记录的通知
如何修复
- 在k8s集群中，所有节点都可以进行时间同步
- 检查所有k8s节点上的NTP同步状态。如果存在任何问题，请解决此问题。
chronyc tracking chronyc sources -v chronyc sourcestats -v
6.警报
| 警报 |
严重级 |
摘要 |
|---|---|---|
| cdlLocalRequestFailure |
关键 |
如果5分钟以上的本地请求成功率低于90%，则触发警报 |
| cdlRemoteConnectionFailure |
关键 |
如果从终端Pod到远程站点的活动连接超过0分钟时间（超过5分钟），则发出警报（仅适用于启用GR的系统） |
| cdlRemoteRequestFailure |
关键 |
如果传入的远程请求成功率小于90%超过5分钟，则触发警报（仅适用于启用了GR的系统） |
| cdlReplicationError |
关键 |
如果cdl-global命名空间中的传出复制请求与本地请求的比率在超过5分钟内低于90%。（仅适用于启用了GR的系统）。升级活动期间预期会出现这些警报，因此您可以忽略它们。 |
| cdlKafkaRemoteReplicationDelay |
关键 |
如果kafka到远程站点的复制延迟超过5分钟，则超过10秒就会发出警报（仅适用于启用了GR的系统） |
| cdlOverloaded — 重大 |
重大 |
如果CDL系统达到配置的容量百分比（默认值为80%），则系统触发警报（仅当启用过载保护功能时） |
| cdlOverloaded — 严重 |
关键 |
如果CDL系统达到配置的容量百分比（默认值为90%），则系统触发警报（仅当启用过载保护功能时） |
| cdlKafkaConnectionFailure |
关键 |
如果CDL索引Pod与kafka断开连接超过5分钟 |
7.最常见的问题
7.1 cdlReplicationError
此警报通常出现在启动运营中心或系统升级期间，尝试为其找到CR，尝试检查CEE出现的警报是否已清除。
7.2 cdlRemoteConnectionFailure & GRPC_Connections_Remote_Site
该说明适用于所有“cdlRemoteConnectionFailure”和“GRPC_Connections_Remote_Site”警报。
对于cdlRemoteConnectionFailure警报：
从CDL终端日志中，我们看到从CDL终端Pod到远程主机的连接已丢失：
2022/01/20 01:36:18.852 [ERROR] [RemoteEndointConnection.go:572] [datastore.ep.session] Connection to remote systemID 2 has been lost
我们可以看到CDL终端Pod尝试连接到远程服务器，但被远程主机拒绝：
2022/01/20 01:37:08.730 [WARN] [checksum.go:836] [datastore.checksum.session] Could not get checksum from remote for systemID: 2, clot: 8, Instance: 2 : rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp 10.10.10.141:8882: connect: connection refused"
2022/01/20 01:37:08.732 [WARN] [checksum.go:836] [datastore.checksum.session] Could not get checksum from remote for systemID: 2, clot: 8, Instance: 1 : rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp 10.10.10.141:8882: connect: connection refused"
2022/01/20 01:37:08.742 [WARN] [checksum.go:836] [datastore.checksum.session] Could not get checksum from remote for systemID: 2, clot: 8, Instance: 1 : rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp 10.10.10.141:8882: connect: connection refused"
2022/01/20 01:37:08.742 [WARN] [checksum.go:836] [datastore.checksum.session] Could not get checksum from remote for systemID: 2, clot: 8, Instance: 2 : rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp 10.10.10.141:8882: connect: connection refused"
2022/01/20 01:37:08.752 [WARN] [checksum.go:836] [datastore.checksum.session] Could not get checksum from remote for systemID: 2, clot: 7, Instance: 2 : rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp 10.10.10.141:8882: connect: connection refused"
2022/01/20 01:37:08.754 [WARN] [checksum.go:836] [datastore.checksum.session] Could not get checksum from remote for systemID: 2, clot: 7, Instance: 1 : rpc error: code = Unavailable desc = connection error: desc = "transport: Error while dialing dial tcp 10.10.10.141:8882: connect: connection refused"
由于远程主机在5分钟内仍无法访问，因此警报引发如下：
alerts history detail cdlRemoteConnectionFailure f5237c750de6
severity critical
type "Processing Error Alarm"
startsAt 2025-01-21T01:41:26.857Z
endsAt 2025-01-21T02:10:46.857Z
source cdl-ep-session-c1-d0-6d86f55945-pxfx9
summary "CDL endpoint connections from pod cdl-ep-session-c1-d0-6d86f55945-pxfx9 and namespace smf-rcdn to remote site reached 0 for longer than 5 minutes"
labels [ "alertname: cdlRemoteConnectionFailure" "cluster: smf-data-rcdn_cee" "monitor: prometheus" "namespace: smf-rcdn" "pod: cdl-ep-session-c1-d0-6d86f55945-pxfx9" "replica: smf-data-rcdn_cee" "severity: critical" ]
annotations [ "summary: CDL endpoint connections from pod cdl-ep-session-c1-d0-6d86f55945-pxfx9 and namespace smf-rcdn to remote site reached 0 for longer than 5 minutes" "type: Processing Error Alarm" ]
在02:10:32:
2022/01/20 02:10:32.702 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
2022/01/20 02:10:38.923 [WARN] [RemoteEndointConnection.go:563] [datastore.ep.session] Cdl status changed - new version 283eb1e86aa9561c653083e6b691c919, old version f81478148c9e1ccb28f3ec0d90ca04e1. Reloading connections
2022/01/20 02:10:38.927 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
2022/01/20 02:10:38.934 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
2022/01/20 02:10:38.934 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
2022/01/20 02:10:38.935 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
CDL远程站点的SMF中存在配置：
cdl remote-site 2
db-endpoint host 10.10.10.141
db-endpoint port 8882
kafka-server 10.10.19.139 10061
exit
kafka-server 10.10.10.140 10061
exit
exit
对于警报GRPC_Connections_Remote_Site:
同样的解释同样适用于“GRPC_Connections_Remote_Site”，因为它也来自同一个CDL终端Pod。
alerts history detail GRPC_Connections_Remote_Site f083cb9d9b8d
severity critical
type "Communications Alarm"
startsAt 2025-01-21T01:37:35.160Z
endsAt 2025-01-21T02:11:35.160Z
source cdl-ep-session-c1-d0-6d86f55945-pxfx9
summary "GRPC connections to remote site are not equal to 4"
labels [ "alertname: GRPC_Connections_Remote_Site" "cluster: smf-data-rcdn_cee" "monitor: prometheus" "namespace: smf-rcdn" "pod: cdl-ep-session-c1-d0-6d86f55945-pxfx9" "replica: smf-data-rcdn_cee" "severity: critical" "systemId: 2" ]
从CDL终端Pod日志，当与远程主机的连接被拒绝时，警报开始：
2022/01/20 01:36:18.852 [ERROR] [RemoteEndointConnection.go:572] [datastore.ep.session] Connection to remote systemID 2 has been lost
当成功连接到远程站点时，警报被清除：
2022/01/20 02:10:32.702 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
2022/01/20 02:10:38.923 [WARN] [RemoteEndointConnection.go:563] [datastore.ep.session] Cdl status changed - new version 283eb1e86aa9561c653083e6b691c919, old version f81478148c9e1ccb28f3ec0d90ca04e1. Reloading connections
2022/01/20 02:10:38.927 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
2022/01/20 02:10:38.934 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
2022/01/20 02:10:38.934 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
2022/01/20 02:10:38.935 [WARN] [RemoteEndointConnection.go:437] [datastore.ep.session] Stream to remote site successful
8.格拉法纳
CDL控制面板是每个SMF部署的一部分。
修订历史记录
| 版本 | 发布日期 | 备注 |
|---|---|---|
1.0 |
04-Oct-2023 |
初始版本 |
反馈