设计一个拥有云原生编排能力、支持多云环境部署、自动化运维、弹性扩缩容、故障自愈等特性,同时提供租户隔离、权限管理、操作审计等企业级能力的高性能、低成本的分布式中间件服务是真挺难的。
Snuba 是一个在 Clickhouse 基础上提供丰富数据模型、快速摄取消费者和查询优化器的服务。以搜索和提供关于 Sentry 事件数据的聚合引擎。
数据完全存储在 Clickhouse 表和物化视图中,它通过输入流(目前只有 Kafka 主题)摄入,可以通过时间点查询或流查询(订阅)进行查询。
文档:
Kubernetes Operator 是一种封装、部署和管理 Kubernetes 应用的方法。我们使用 Kubernetes API(应用编程接口)和 kubectl 工具在 Kubernetes 上部署并管理 Kubernetes 应用。
Altinity:ClickHouse Operator 业界领先开源提供商。
当然这种多租户隔离的 ClickHouse 中间件 PaaS 云平台,公司或云厂商几乎是不开源的。
云厂商(青云)基于 altinity-clickhouse-operator 定制的。对于快速部署生产集群做了些优化。
VKE K8S Cluster,Vultr
托管集群(v1.23.14)
Kubesphere v3.3.1 集群可视化管理,全栈的 Kubernetes 容器云 PaaS 解决方案。
Longhorn 1.14,Kubernetes 的云原生分布式块存储。
这里我们使用 RadonDB 定制的 Operator。
values.operator.yaml
定制如下两个参数:
watchAllNamespaces: true
enablePrometheusMonitor: true
helm 部署 operator:
cd vip-k8s-paas/10-cloud-native-clickhouse
helm install clickhouse-operator ./clickhouse-operator -f values.operator.yaml -n kube-system
kubectl -n kube-system get po | grep clickhouse-operator
kubectl -n kube-system get svc | grep clickhouse-operator
kubectl api-resources | grep clickhouse
这里我们使用 RadonDB 定制的 clickhouse-cluster helm charts。
快速部署 2 shards + 2 replicas + 3 zk nodes 的集群。
values.cluster.yaml
定制:
clickhouse:
clusterName: snuba-clickhouse-nodes
shardscount: 2
replicascount: 2
…
zookeeper:
install: true
replicas: 3
helm 部署 clickhouse-cluster:
kubectl create ns cloud-clickhouse
helm install clickhouse ./clickhouse-cluster -f values.cluster.yaml -n cloud-clickhouse
kubectl get po -n cloud-clickhouse
使用如下命令,将 shardsCount
改为 3
:
kubectl edit chi/clickhouse -n cloud-clickhouse
查看 pods:
kubectl get po -n cloud-clickhouse
发现多出 chi-clickhouse-snuba-ck-nodes-2-0-0
与 chi-clickhouse-snuba-ck-nodes-2-1-0
。 分片与副本已自动由 Operator
新建。
我们进入 Pod, 使用原生命令行客户端 clickhouse-client
连接。
kubectl exec -it chi-clickhouse-snuba-ck-nodes-0-0-0 -n cloud-clickhouse -- bash
kubectl exec -it chi-clickhouse-snuba-ck-nodes-0-1-0 -n cloud-clickhouse -- bash
kubectl exec -it chi-clickhouse-snuba-ck-nodes-1-0-0 -n cloud-clickhouse -- bash
kubectl exec -it chi-clickhouse-snuba-ck-nodes-1-1-0 -n cloud-clickhouse -- bash
kubectl exec -it chi-clickhouse-snuba-ck-nodes-2-0-0 -n cloud-clickhouse -- bash
kubectl exec -it chi-clickhouse-snuba-ck-nodes-2-1-0 -n cloud-clickhouse -- bash
我们直接通过终端分别进入这 6 个 pod。然后进行测试:
clickhouse-client --multiline -u username -h ip --password passowrd
# clickhouse-client -m
查看 system.clusters
select * from system.clusters;
2.创建名为 test
的数据库
create database test on cluster 'snuba-ck-nodes';
# 删除:drop database test on cluster 'snuba-ck-nodes';
在各个节点查看,都已存在 test
数据库。
show databases;
在集群中各个节点 test
数据库中创建 t_local
本地表,采用 ReplicatedMergeTree
表引擎,接受两个参数:
zoo_path
— zookeeper 中表的路径,针对表同一个分片的不同副本,定义相同路径。
replica_name
— zookeeper 中表的副本名称
CREATE TABLE test.t_local on cluster 'snuba-ck-nodes'
(
EventDate DateTime,
CounterID UInt32,
UserID UInt32
)
ENGINE = ReplicatedMergeTree('/clickhouse/tables/{shard}/test/t_local', '{replica}')
PARTITION BY toYYYYMM(EventDate)
ORDER BY (CounterID, EventDate, intHash32(UserID))
SAMPLE BY intHash32(UserID);
macros
)占位符:建表语句参数包含的宏替换占位符(如:{replica}
)。会被替换为配置文件里 macros 部分的值。
查看集群中 clickhouse 分片&副本节点 configmap
:
kubectl get configmap -n cloud-clickhouse | grep clickhouse
NAME DATA AGE
chi-clickhouse-common-configd 6 20h
chi-clickhouse-common-usersd 6 20h
chi-clickhouse-deploy-confd-snuba-ck-nodes-0-0 2 20h
chi-clickhouse-deploy-confd-snuba-ck-nodes-0-1 2 20h
chi-clickhouse-deploy-confd-snuba-ck-nodes-1-0 2 20h
chi-clickhouse-deploy-confd-snuba-ck-nodes-1-1 2 20h
chi-clickhouse-deploy-confd-snuba-ck-nodes-2-0 2 19h
chi-clickhouse-deploy-confd-snuba-ck-nodes-2-1 2 19h
查看节点配置值:
kubectl describe configmap chi-clickhouse-deploy-confd-snuba-ck-nodes-0-0 -n cloud-clickhouse
CREATE TABLE test.t_dist on cluster 'snuba-ck-nodes'
(
EventDate DateTime,
CounterID UInt32,
UserID UInt32
)
ENGINE = Distributed('snuba-ck-nodes', test, t_local, rand());
# drop table test.t_dist on cluster 'snuba-ck-nodes';
这里,Distributed 引擎的所用的四个参数:
snuba-ck-nodes
)test
)t_local
)CounterID/rand()
)查看相关表,如:
use test;
show tables;
# t_dist
# t_local
通过分布式表插入几条数据:
# 插入
INSERT INTO test.t_dist VALUES ('2022-12-16 00:00:00', 1, 1),('2023-01-01 00:00:00',2, 2),('2023-02-01 00:00:00',3, 3);
任一节点查询数据:
select * from test.t_dist;
在我们迁移到 Kubernetes Operator 之前,我们先拆解与分析下 sentry-charts 中自带的 clickhouse & zookeeper charts。
非官方 Sentry Helm Charts:
他的 Chart.yaml
如下:
apiVersion: v2
appVersion: 22.11.0
dependencies:
- condition: sourcemaps.enabled
name: memcached
repository: https://charts.bitnami.com/bitnami
version: 6.1.5
- condition: redis.enabled
name: redis
repository: https://charts.bitnami.com/bitnami
version: 16.12.1
- condition: kafka.enabled
name: kafka
repository: https://charts.bitnami.com/bitnami
version: 16.3.2
- condition: clickhouse.enabled
name: clickhouse
repository: https://sentry-kubernetes.github.io/charts
version: 3.2.0
- condition: zookeeper.enabled
name: zookeeper
repository: https://charts.bitnami.com/bitnami
version: 9.0.0
- alias: rabbitmq
condition: rabbitmq.enabled
name: rabbitmq
repository: https://charts.bitnami.com/bitnami
version: 8.32.2
- condition: postgresql.enabled
name: postgresql
repository: https://charts.bitnami.com/bitnami
version: 10.16.2
- condition: nginx.enabled
name: nginx
repository: https://charts.bitnami.com/bitnami
version: 12.0.4
description: A Helm chart for Kubernetes
maintainers:
- name: sentry-kubernetes
name: sentry
type: application
version: 17.9.0
这个 sentry-charts 将所有中间件 helm charts 耦合依赖在一起部署,不适合 sentry 微服务 & 中间件集群扩展。更高级的做法是每个中间件拥有定制的 Kubernetes Operator(如:clickhouse-operator
) & 独立的 K8S 集群,形成中间件 PaaS 平台对外提供服务。
这里我们拆分中间件 charts 到独立的 namespace 或单独的集群运维。设计为:
cloud-zookeeper-paas
cloud-clickhouse-paas
这里 zookeeper chart 采用的是 bitnami/zookeeper,他的仓库地址如下:
创建命名空间:
kubectl create ns cloud-zookeeper-paas
简单定制下 values.yaml
:
metrics:
containerPort: 9141
enabled: true
….
….
service:
annotations: {}
clusterIP: ""
disableBaseClientPort: false
externalTrafficPolicy: Cluster
extraPorts: []
headless:
annotations: {}
publishNotReadyAddresses: true
loadBalancerIP: ""
loadBalancerSourceRanges: []
nodePorts:
client: ""
tls: ""
ports:
client: 2181
election: 3888
follower: 2888
tls: 3181
sessionAffinity: None
type: ClusterIP
注意:在使用支持外部负载均衡器的云提供商的服务时,需设置 Sevice 的 type 的值为 "LoadBalancer", 将为 Service 提供负载均衡器。来自外部负载均衡器的流量将直接重定向到后端 Pod 上,不过实际它们是如何工作的,这要依赖于云提供商。
helm 部署:
helm install zookeeper ./zookeeper -f values.yaml -n cloud-zookeeper-paas
集群内,可使用 zookeeper.cloud-zookeeper-paas.svc.cluster.local:2181
对外提供服务。
zkCli 连接 ZooKeeper:
export POD_NAME=$(kubectl get pods --namespace cloud-zookeeper-paas -l "app.kubernetes.io/name=zookeeper,app.kubernetes.io/instance=zookeeper,app.kubernetes.io/component=zookeeper" -o jsonpath="{.items[0].metadata.name}")
kubectl -n cloud-zookeeper-paas exec -it $POD_NAME -- zkCli.sh
[zk: localhost:2181(CONNECTED) 0] ls /
[zookeeper]
[zk: localhost:2181(CONNECTED) 1] ls /zookeeper
[config, quota]
[zk: localhost:2181(CONNECTED) 2] quit
查看 zoo.cfg
kubectl -n cloud-zookeeper-paas exec -it $POD_NAME -- cat /opt/bitnami/zookeeper/conf/zoo.cfg
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/bitnami/zookeeper/data
clientPort=2181
maxClientCnxns=60
#
#
#
autopurge.snapRetainCount=3
autopurge.purgeInterval=0
#
metricsProvider.className=org.apache.zookeeper.metrics.prometheus.PrometheusMetricsProvider
#metricsProvider.httpHost=0.0.0.0
metricsProvider.httpPort=9141
metricsProvider.exportJvmInfo=true
preAllocSize=65536
snapCount=100000
maxCnxns=0
reconfigEnabled=false
quorumListenOnAllIPs=false
4lw.commands.whitelist=srvr, mntr, ruok
maxSessionTimeout=40000
admin.serverPort=8080
admin.enableServer=true
server.1=zookeeper-0.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local:2888:3888;2181
server.2=zookeeper-1.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local:2888:3888;2181
server.3=zookeeper-2.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local:2888:3888;2181
这里 clickhouse chart 采用的是 sentry-kubernetes/charts 自己维护的一个版本:
yandex/clickhouse-server:20.8.19.4
。这个自带的 clickhouse-charts 存在些问题,Service 部分需简单修改下允许配置 "type:LoadBalancer" or "type:NodePort"。
注意:在使用支持外部负载均衡器的云提供商的服务时,需设置 Sevice 的 type 的值为 "LoadBalancer", 将为 Service 提供负载均衡器。来自外部负载均衡器的流量将直接重定向到后端 Pod 上,不过实际它们是如何工作的,这要依赖于云提供商。
创建命名空间:
kubectl create ns cloud-clickhouse-paas
简单定制下 values.yaml
:
注意上面 zoo.cfg
的 3 个 zookeeper 实例的地址:
server.1=zookeeper-0.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local:2888:3888;2181
server.2=zookeeper-1.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local:2888:3888;2181
server.3=zookeeper-2.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local:2888:3888;2181
# 修改 zookeeper_servers
clickhouse:
configmap:
zookeeper_servers:
config:
- hostTemplate: 'zookeeper-0.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local'
index: clickhouse
port: "2181"
- hostTemplate: 'zookeeper-1.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local'
index: clickhouse
port: "2181"
- hostTemplate: 'zookeeper-2.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local'
index: clickhouse
port: "2181"
enabled: true
operation_timeout_ms: "10000"
session_timeout_ms: "30000"
# 暴露下 prometheus 监控所需的服务
metrics:
enabled: true
当然这里也可以不用 Headless Service,因为是同一个集群的不同 namespace 的内部访问,所以也可简单填入 ClusterIP 类型 Sevice:
# 修改 zookeeper_servers
clickhouse:
configmap:
zookeeper_servers:
config:
- hostTemplate: 'zookeeper.cloud-zookeeper-paas.svc.cluster.local'
index: clickhouse
port: "2181"
enabled: true
operation_timeout_ms: "10000"
session_timeout_ms: "30000"
# 暴露下 prometheus 监控所需的服务
metrics:
enabled: true
helm 部署:
helm install clickhouse ./clickhouse -f values.yaml -n cloud-clickhouse-paas
连接 clickhouse
kubectl -n cloud-clickhouse-paas exec -it clickhouse-0 -- clickhouse-client --multiline --host="clickhouse-1.clickhouse-headless.cloud-clickhouse-paas"
验证集群
show databases;
select * from system.clusters;
select * from system.zookeeper where path = '/clickhouse';
kubectl get configmap -n cloud-clickhouse-paas | grep clickhouse
clickhouse-config 1 28h
clickhouse-metrica 1 28h
clickhouse-users 1 28h
config.xml
)<yandex>
<path>/var/lib/clickhouse/</path>
<tmp_path>/var/lib/clickhouse/tmp/</tmp_path>
<user_files_path>/var/lib/clickhouse/user_files/</user_files_path>
<format_schema_path>/var/lib/clickhouse/format_schemas/</format_schema_path>
<include_from>/etc/clickhouse-server/metrica.d/metrica.xml</include_from>
<users_config>users.xml</users_config>
<display_name>clickhouse</display_name>
<listen_host>0.0.0.0</listen_host>
<http_port>8123</http_port>
<tcp_port>9000</tcp_port>
<interserver_http_port>9009</interserver_http_port>
<max_connections>4096</max_connections>
<keep_alive_timeout>3</keep_alive_timeout>
<max_concurrent_queries>100</max_concurrent_queries>
<uncompressed_cache_size>8589934592</uncompressed_cache_size>
<mark_cache_size>5368709120</mark_cache_size>
<timezone>UTC</timezone>
<umask>022</umask>
<mlock_executable>false</mlock_executable>
<remote_servers incl="clickhouse_remote_servers" optional="true" />
<zookeeper incl="zookeeper-servers" optional="true" />
<macros incl="macros" optional="true" />
<builtin_dictionaries_reload_interval>3600</builtin_dictionaries_reload_interval>
<max_session_timeout>3600</max_session_timeout>
<default_session_timeout>60</default_session_timeout>
<disable_internal_dns_cache>1</disable_internal_dns_cache>
<query_log>
<database>system</database>
<table>query_log</table>
<partition_by>toYYYYMM(event_date)</partition_by>
<flush_interval_milliseconds>7500</flush_interval_milliseconds>
</query_log>
<query_thread_log>
<database>system</database>
<table>query_thread_log</table>
<partition_by>toYYYYMM(event_date)</partition_by>
<flush_interval_milliseconds>7500</flush_interval_milliseconds>
</query_thread_log>
<distributed_ddl>
<path>/clickhouse/task_queue/ddl</path>
</distributed_ddl>
<logger>
<level>trace</level>
<log>/var/log/clickhouse-server/clickhouse-server.log</log>
<errorlog>/var/log/clickhouse-server/clickhouse-server.err.log</errorlog>
<size>1000M</size>
<count>10</count>
</logger>
</yandex>
metrica.xml
)<yandex>
<zookeeper-servers>
<node index="clickhouse">
<host>zookeeper-0.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local</host>
<port>2181</port>
</node>
<node index="clickhouse">
<host>zookeeper-1.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local</host>
<port>2181</port>
</node>
<node index="clickhouse">
<host>zookeeper-2.zookeeper-headless.cloud-zookeeper-paas.svc.cluster.local</host>
<port>2181</port>
</node>
<session_timeout_ms>30000</session_timeout_ms>
<operation_timeout_ms>10000</operation_timeout_ms>
<root></root>
<identity></identity>
</zookeeper-servers>
<clickhouse_remote_servers>
<clickhouse>
<shard>
<replica>
<internal_replication>true</internal_replication>
<host>clickhouse-0.clickhouse-headless.cloud-clickhouse-paas.svc.cluster.local</host>
<port>9000</port>
<user>default</user>
<compression>true</compression>
</replica>
</shard>
<shard>
<replica>
<internal_replication>true</internal_replication>
<host>clickhouse-1.clickhouse-headless.cloud-clickhouse-paas.svc.cluster.local</host>
<port>9000</port>
<user>default</user>
<compression>true</compression>
</replica>
</shard>
<shard>
<replica>
<internal_replication>true</internal_replication>
<host>clickhouse-2.clickhouse-headless.cloud-clickhouse-paas.svc.cluster.local</host>
<port>9000</port>
<user>default</user>
<compression>true</compression>
</replica>
</shard>
</clickhouse>
</clickhouse_remote_servers>
<macros>
<replica from_env="HOSTNAME"></replica>
<shard from_env="SHARD"></shard>
</macros>
</yandex>
users.xml
)<yandex>
</yandex>
我们简单修改 values.yml
clickhouse:
enabled: false
zookeeper:
enabled: false
externalClickhouse
externalClickhouse:
database: default
host: "clickhouse.cloud-clickhouse-paas.svc.cluster.local"
httpPort: 8123
password: ""
singleNode: false
clusterName: "clickhouse"
tcpPort: 9000
username: default
注意:
这里只是简单的集群内部接入 1 个多节点分片集群,而 Snuba 系统的设计是允许你接入多个 ClickHouse 多节点多分片多副本集群,将多个 Schema 分散到不同的集群,从而实现超大规模吞吐。因为是同一个集群的不同 namespace 的内部访问,所以这里简单填入类型为 ClusterIP Sevice 即可。
注意这里 singleNode
要设置成 false
。因为我们是多节点,同时我们需要提供 clusterName
:
源码分析:
这将用于确定:
以及确定来使用不同的 ClickHouse Table Engines 等。
当然,ClickHouse 本身是一个单独的技术方向,这里就不展开讨论了。
helm install sentry ./sentry -f values.yaml -n sentry
kubectl -n cloud-clickhouse-paas exec -it clickhouse-0 -- clickhouse-client --multiline --host="clickhouse-1.clickhouse-headless.cloud-clickhouse-paas"
show databases;
show tables;
select * from system.zookeeper where path = '/clickhouse';
独立部署多套 VKE LoadBlancer+ VKE K8S Cluster + ZooKeeper-Operator + ClickHouse-Operator,分散 Schema 到不同的集群以及多节点分片。
关于针对 ClickHouse 集群各个分片、副本之间的读写负载均衡、连接池等问题。Snuba 在系统设计、代码层面部分就已经做了充分的考虑以及优化。
关于 ClickHouse Operator 独立的多个云原生编排集群以及 Snuba 系统设计等高级部分会在 VIP 专栏直播课单独讲解。
手机扫一扫
移动阅读更方便
你可能感兴趣的文章