Scaling a Multi-Tenant k8s Cluster in a Telco Pablo Moncada October 28, 2020 About MasMovil group ● 4th telecom company in Spain ● Provides voice and broadband services to +12M customers ● Several Services +3k CPU +2k Mem +5TB Nodes +300 kube-proxy replacement NetworkPolicy logging Multi-cluster DNS Aware NetworkPolicy Increased Istio security External Services TLS visibility Performance

PolarDB-X 如 何 做 生 态兼 容 好的 MySQL 兼容性可以做到什么程度 胡中泉（舟济） 阿里云数据库解决方案架构师为什么要兼容 MySQL 01 The longer you look back, the farther you can look forward.也从阿里巴巴的“去IOE”运动说起 业务驱动下的分布式技术实践之路 5月17日，支付宝最后一台小型 机下线标志去IOE落下帷幕 反推出了天价账单也谈所谓的“中间件” 中间件只是起点，PolarDB-X 可能是离终点最近的那个 对近十年的探索以及五年的上云 经验进行重新思考，面向未来设 从运维视角实现计算存储一体化 计新架构 产品形态 基于MySQL XA实现分布式事务 基于外部组件进行扩容 支持扩容 分布式事务 一体化尝试 Review 2.0 开源 2016 2017 2018 2019 2020 2021 2021年10月20日，云栖大会宣 it probably is a duck.业务系统 上游 Single Source of Truth 问题背景 数据孤岛怎么办？ 下游MySQL 怎么做 Binlog Maxwell DebeziumPolarDB-X 完全兼容 MySQL Binlog 可行性 • 多节点产生多个增量事件队列 • 不同队列中事件之间的顺序 • 分布式事务完整性 • DDL 引起的多 Schema

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1804 4 Monitor MySQL with Zabbix agent 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . code points (instead of 255 in previous versions). A longer value will be truncated. Note that on MySQL, the effective maximum length in characters will be less if the returned value contains multibyte What’s new in Zabbix 6.4.5 Templates New templates are available: • AWS ECS Cluster by HTTP (along with its Serverless Cluster version) • Cisco SD-WAN by HTTP • OpenStack by HTTP, which includes OpenStack

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1862 4 Monitor MySQL with Zabbix agent 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . connector - statistics for connectors with the largest queue ha_status Log high availability (HA) cluster status. ha_remove_node=target Remove the high availability (HA) node specified by its name or ID Zabbix high availability mode multiple Zabbix servers are run as nodes in a cluster. While one Zabbix server in the cluster is active, others are on standby, ready to take over if necessary. 24 Switching

setup. 5 What’s new in Zabbix 6.0.0 See breaking changes for this version. High availability cluster for Zabbix server The new version comes with a native high availability solution for Zabbix server standby, ready to take over in case the current node is stopped or fails. See also: High availability cluster. Updated service monitoring Several updates have been made to the monitoring of services. Service performance. Starting from Zabbix 6.0, the following database versions are officially supported: • MySQL/Percona 8.0.X • MariaDB 10.5.X - 10.6.X • PostgreSQL 13.X - 14.X • Oracle 19c - 21c • TimescaleDB

custom attribute values. The items vmware.hv.discovery[], used for hypervisor discovery, and vmware.cluster.discovery[], used for clusters discovery, now return information about resource pools. Active checks monitoring New items are now available for: • Monitoring VMware alarms: vmware.alarms.get[], vmware.cluster.alarms.get[], vmware.datastore.alarms.get[], vmware.dc.alarms.get[], vmware.hv.alarms.get[], vmware VMware components: vmware.cluster.tags.get[], vmware.datastore.tags.get[], vmware.dc.tags.get[], vmware.hv.tags.get[], vmware.vm.tags.get[]. • Returning property values: vmware.cluster.property[], vmware.datastore

your datacentre, public or private. Whether you want to deploy an OpenStack cloud, a Kubernetes cluster or a 50,000-node render farm, Ubuntu Server delivers the best value scale-out performance available HA - Migrate from crmsh to pcs 2 Databases 3 databases-introduction Introduction 3 databases-mysql MySQL 3 databases-postgresql PostgreSQL 2 Monitoring 3 logging-monitoring-alerting Logging, Monitoring create -f qcow2 disk-image.qcow2 10G Formatting 'disk-image.qcow2', fmt=qcow2 size=10737418240 cluster_size=65536 lazy_refcounts=off refcount_bits=16 qemu-img info disk-image.qcow2 image: disk-image

Maximized node performance Application cluster Storage cluster (multi-node/multi-controller) /A /A /B /C /D /E /F Multipath NFS/RDMA or NFS/TCP Maximized cluster performance in client-server model improving the application performance. Repositories https://gitee.com/openeuler/wayca-scheduler • Cluster and NUMA scheduling domains are established based on the hardware topology, and the scheduler supports HybridSched is a full-stack solution for hybrid deployment of VMs, covering enhanced OpenStack cluster scheduling, new single-node QoS management component Skylark, and kernel-mode basic resource isolation

Parameter structure for Zabbix agent 2 plugins has been updated: • For Ceph, Docker, Memcached, MySQL, Oracle, and Redis plugins, a Uri can now only be specified as a named session parameter. • For SharePont Microsoft SharePoint by HTTP - see setup instructions for HTTP templates. MongoDB • MongoDB cluster by Zabbix agent 2 • MongoDB node by Zabbix agent 2 See setup instructions for Zabbix agent 2 templates monitoring of TiDB server of TiDB cluster via HTTP; • TiDB PD by HTTP - monitoring of PD server of TiDB cluster via HTTP; • TiDB TiKV by HTTP - monitor TiKV server of TiDB cluster via HTTP. See setup instructions

item, referencing the current host sum(last_foreach(/*/vfs.fs.size[/,total]?[group=”MySQL Servers”])) grpsum[”MySQL Servers”,”vfs.fs.size[/,total]”,last] Aggregate calculation (formerly, aggregate performance. Starting from Zabbix 5.4, the following database versions are officially supported: • MySQL/Percona 5.7.28-8.0.X • MariaDB 10.0.37-10.5.X • PostgreSQL 10.9-13.X • Oracle 12.1.0.2 - 19c • setupfor more information. Items VMware cluster performance counters A new vmware.cl.perfcounter[,,,] item allows to retrieve VMware cluster performance counter metrics. JMX items