译者: bzhaoopenstack
作者: Amit Dattatray Khandekar
原文链接: https://amitdkhan-pg.blogspot.com/2021/04/a-quick-sanity-testing-of-parallel.html

PG在很久之前就已经支持并行查询，但是对于Arm64平台它是否有完备性是需要考究的。有很大的可能是还没有人真正完备的测试过，下面我们通过测试结果来看看它在ARM上的表现，同时理解并行查询的真正含义。

译者: bzhaoopenstack
作者: Amit Dattatray Khandekar
原文链接: https://amitdkhan-pg.blogspot.com/2020/08/need-for-external-compression-methods.html

数据库中的压缩算法对DBA和后续的维护成本至关重要，PG社区上游正在讨论如何为用户提供自定义native压缩算法，包括软硬件压缩，对于DB产业价值很高。

作者: bzhaoopenstack

2020年是非常令人难忘的一年，2020年是我们包含的无数个首次，居家隔离，在家办公，千辛万苦从老家赶往单位上班，从艰难的上班路程一直到现在越来越趋于正常的环境。回想这一整年，我们从今年年后开始组建团队在社区上游进行鲲鹏优化相关的工作，一开始我们只有两个人，到现在的初具规模，伙伴们从一个ARM优化小白成长到能够独挡一面的先驱者，这一年的经历相信会在我们的人生当中留下浓重的一笔。

Author:zhaorenhai

MariaDB’s connect storage engine is an exciting storage engine.

The reason why it’s exciting is that the storage engine can do so much. Except for OLTP field, it is almost omnipotent. It can directly use SQL to query external files. It supports various common text files and log files. As long as you tell CONNECT the record format, you can use SQL to query and analyze. It also supports common file formats such as JSON, XML and INI. In addition to supporting local files, remote data can also be queried directly through the rest interface. For local files, in addition to query, it also supports insert, delete and other functions. It can also realize the dblink function in Oracle, that is, it can remotely query the tables of another database, even different types of databases, including Mysql, PostgreSQL, Oracle, etc., and even MongoDB of NoSQL type. Of course, on this basis, it is easier to import external files into the database, and it is more powerful and flexible than mysqlimport, because it can realize various format conversion and filtering. The same is true for exporting data to an external file. It also supports querying compressed files (currently only in ZIP format). Even binary file queries and updates are supported. The engine also supports virtual table which functions similar to Oracle’s dual. It can also support a variety of functions, such as single row to multi row, row column conversion, multi table mapping to a table.

In a word, with the CONNECT storage engine, everything can be transformed into SQL, and SQL can be transformed into everything.

Next, let’s take a look at the performance of the connect storage engine on the arm64 platform.

作者: zhaorenhai

MariaDB的CONNECT存储引擎是一个让人感到兴奋的存储引擎。

之所以说让人感到兴奋，是因为这个存储引擎能做的事实在是太多了。除了不适合用于OLTP领域之外，几乎是上天入地，无所不能。它可以直接用SQL去查询外部文件，支持各种普通的文本文件，日志文件，只要告诉CONNECT固定的文件格式，就可以用SQL去查询，去分析。它还支持JSON，XML，INI这些常见的文件格式。除了支持本地的文件，还可以通过REST接口直接查询远程的数据。对于本地的文件，除了查询，还支持插入，删除等功能。它还可以实现Oracle里面的dblink功能，也就是可以远程查询另外一个数据库的表，甚至是不同类型的数据库，包括MySQL，PostgreSQL，Oracle等等，甚至可以是NoSQL类型的MongoDB。当然所谓的跨源查询，多库查询，更是不在话下。当然在此基础上，将外部文件导入到数据库就更简单了，而且比mysqlimport更强大，更灵活，因为可以实现各种各样的格式转换和过滤。将数据导出到外部文件也是一样。它还支持对压缩文件的查询(目前只有zip格式)。甚至对于二进制格式的文件查询和更新都支持。。这个引擎还支持类似于Oracle的dual的虚拟表功能。还可以支持单行转多行，行列转换，多表映射成一个表等各种各样的功能。

总之有了这个CONNECT存储引擎，可以说，一切都可以转化成SQL，SQL也可以转化成一切。

接下来我们就看看CONNECT存储引擎在arm64平台上的表现。

Author: zhaorenhai

This article attempts to explore some useful performance monitoring tools on the arm64 platform that can be used to monitor MariaDB databases.

Author: zhaorenhai

This article attempts to explore how several top programming languages connect to MariaDB database on the arm64 platform, and test whether the addition, deletion, modification, and query functions are normal. Include the following languages: C, Java, Node.js, Python, Go, Rust, PHP

Author: zhaorenhai

This article plans to test the operation of MariaDB’s backup and recovery function on the arm64 platform, and provide a reference for those interested in deploying MariaDB on the arm64 platform.

The test platform chooses the 8C16G Kunpeng virtual machine on Huawei Cloud, the OS is Ubuntu 18.04, and MariaDB we plan to choose the latest version on github for testing.

Author: zhaorenhai

This article attempts to deploy MariaDB clusters on x86 and arm64 platforms to see if the deployment can be successful.
The deployment environment is carried out on HUAWEI CLOUD. The OS is openEuler 20.03 version and MariaDB version is 10.3.9.
In the official MariaDB documentation, there are two types of high-availability environment deployment methods. The simpler one is replication, and then the Galera cluster.
Let’s try these two high-availability environments separately, whether they support mixed deployment on x86 and arm64 platforms.

作者: wangxiyuan

本文介绍自旋锁的概念、实现以及优化