vertica_将企业vertica转换为eon

vertica

If you are considering this move, Congratulations! You are headed in the right direction. Eon mode is a giant leap forward for Vertica.

如果您正在考虑此举，那么恭喜！ 您朝着正确的方向前进。 Eon模式对于Vertica而言是巨大的飞跃。

Eon advances Vertica cluster architecture in several major ways.

Eon通过几种主要方式改进了Vertica集群体系结构。

Firstly, Enterprise mode clusters have data stored in the actual nodes. You’ll have to do some capacity planning. It’s very important to accurately estimate the node size (memory, cpu, disk etc.) that is enough to store and retrieve data in the desired time. Alternatively, Eon Mode stores data in a shared object store called communal storage. This greatly simplifies planning for hardware resources. This is especially useful in cloud environments.

首先，企业模式群集将数据存储在实际节点中。 您必须进行一些容量规划。 准确估计足以在所需时间内存储和检索数据的节点大小(内存，cpu，磁盘等)非常重要。 或者，Eon模式将数据存储在称为公共存储的共享对象存储中。 这大大简化了硬件资源的计划。 这在云环境中尤其有用。

Secondly, data replication tactics in Enterprise mode greatly depend on your K-safety setting. This too requires smart capacity planning and changes performance depending on the number of nodes in your Vertica cluster. The higher the K-safety number, the higher the number of nodes involved in replication which leads to more maintenance.

其次，企业模式下的数据复制策略在很大程度上取决于您的K安全设置。 这也需要智能容量规划并根据Vertica群集中的节点数来更改性能。 K安全数越高，复制涉及的节点数越多，从而导致更多的维护。

With regard to elastic scalability, database size has virtually no limit in Eon. You can keep adding as much data as you wish without having to add new nodes. Data replication in AWS is inherent in S3 which lowers risk with Vertica data. Therefore your EBS costs are much lower than they ordinarily would be. All the maintenance and complexity associated with data replication disappears immediately in Eon.

关于弹性可伸缩性，数据库大小实际上在Eon中没有限制。 您可以不断添加所需的数据，而不必添加新的节点。 AWS中的数据复制是S3固有的功能，可降低Vertica数据的风险。 因此，您的EBS成本比通常要低得多。 与数据复制相关的所有维护和复杂性在Eon中立即消失。

There will always be a cost to make Enterprise Vertica available even when you don’t need it during certain times of the day or night. You can make some limited adjustment but it’s not very elastic in nature especially when it comes to scaling down. You are always paying since you must keep minimum number of nodes at any point of time to maintain your K-safety.

即使您在白天或晚上的某些时间不需要Enterprise Vertica，也总是要付出代价。 您可以进行一些有限的调整，但是它的弹性并不是很大，尤其是在缩小比例方面。 您始终在付款，因为您必须在任何时间保持最少数量的节点以维护K安全。

Eon shines in this area as well. You can literally scale down to zero nodes during quiet periods.

Eon也在这个领域发光。 您可以在安静期间按比例缩小到零节点。

Another beautiful feature that you would fall in love with is optimization of resources in a way that querying and writing data is without impacting load on its performance. Workloads in Eon can be isolated to a subset of nodes.

您可能会喜欢的另一个漂亮功能是资源优化，即查询和写入数据不会影响其性能负载。 Eon中的工作负载可以隔离到一部分节点。

Migrating existing Enterprise clusters to Eon can be challenging. However the advantages are worth it. With Eon there is far less risk of dealing with downtime issues that are typical of those sometimes found on an Enterprise mode cluster node. You may have dealt with some of those issues in the past. If so, you know that resolving them required spending a lot of time trying to troubleshoot, cluster recreation, copy from DR clusters, and general frustration. We have automated many of those procedures in ElasticDW, but EON reduces the risk and overhead. In addition, for some of our clusters we were able to reduce the number of Vertica nodes with no loss of performance or availability. In one case we were able to reduce our AWS costs down to a third — $2400/month to $800/month.

将现有企业集群迁移到Eon可能会面临挑战。 但是优点是值得的。 借助Eon，处理有时在企业模式群集节点上常见的停机时间的风险大大降低。 您过去可能已经处理了其中一些问题。 如果是这样，您就知道解决这些问题需要花费大量时间来尝试进行故障排除，集群重新创建，从灾难恢复集群进行复制以及普遍沮丧。 我们已经在ElasticDW中使许多过程自动化，但是EON降低了风险和开销。 此外，对于我们的某些集群，我们能够减少Vertica节点的数量，而不会降低性能或可用性。 在一种情况下，我们能够将AWS成本降低至三分之一-从$ 2400 /月降至$ 800 /月 。

That specific installation was a three node Enterprise cluster converted into a one node Eon cluster. In this particular use case we had two clusters in an decoupled active-active configuration, and reducing k-safety was not a concern — we can always restore from the other cluster if one goes down (or rebuild from scratch from the data lake using our Pitbull architecture).

该特定安装是将三节点企业群集转换为一个节点Eon群集。 在这个特殊的用例中，我们有两个处于主动/主动分离配置的集群，而降低k安全性不是问题-如果一个集群出现故障，我们总是可以从另一个集群中恢复(或者使用我们的数据库从零开始进行重建) 比特犬架构)。

Here are some tips on how to make the conversion to Eon mode simpler. We have executed the Enterprise to Eon transition for most of our customers’ production workloads, and countless times on our own. In an upcoming release of ElasticDW, we will make this conversion a one click step.

以下是有关如何简化向Eon模式转换的一些提示。 对于大多数客户的生产工作负载，我们已经执行了从企业到Eon的过渡，而我们自己经历了无数次。 在即将发布的ElasticDW版本中，我们将使这一转换成为一键式步骤。

To set context:

设置上下文：

Source: Existing Vertica database in Enterprise mode. (3 node cluster)Target: New Vertica database in Eon mode. (1 node cluster)

来源 ：企业模式下的现有Vertica数据库。 (3节点群集) 目标 ：Eon模式下的新Vertica数据库。 (1个节点群集)

这些是将Vertica从Enterprise转换为Eon的基本步骤： (These are the basic steps to convert Vertica from Enterprise to Eon:)

1. Create Eon cluster with one node (if you have 3 node existing Vertica cluster,1 node Eon cluster as per my findings, this would give you similar performance as before)
   创建一个节点的Eon群集(如果根据我的发现，如果现有3个节点的Vertica群集，1个节点的Eon群集，这将为您提供与以前类似的性能)

2. Get your complete DDL from source using following sql and execute in target.

   使用以下sql从源获取完整的DDL，并在目标中执行。

   SELECT export_objects('', '', FALSE);

   SELECT export_objects('', '', FALSE);

3. Create a script using system tables to copy all users, roles, grant roles/other permissions, etc…

   使用系统表创建脚本来复制所有用户，角色，授予角色/其他权限等…

   Here is an example:

   这是一个例子：

   SELECT 'create user ' || user_name || ' resource pool ' || resource_pool || ';' from v_catalog.users where user_name not in ('public','dbduser','dbadmin','pseudosuperuser','sysmonitor','vertica_admin','system_monitor');Once you run this in source, get the output and run it in target.

   SELECT 'create user ' || user_name || ' resource pool ' || resource_pool || ';' from v_catalog.users where user_name not in ('public','dbduser','dbadmin','pseudosuperuser','sysmonitor','vertica_admin','system_monitor'); 在源代码中运行此命令后，获取输出并在目标中运行它。

   Note: This is just to give you an idea, this is not even 1% of the actually script.

   注意：这只是为了给您一个想法，这甚至不到实际脚本的1％。

4. Create script in source to load all the data in the target.

   在源代码中创建脚本以加载目标中的所有数据。

   Here is the script:

   这是脚本：

SELECT 'CONNECT TO VERTICA my_database USER dbadmin PASSWORD ''XXX'' ON ''SOURCE_IP_ADDRESS'',5433;' union all (select 'select ''copying....' || schema_name || '.' || table_name || '''; ' || 'COPY ' || schema_name || '.' || table_name || ' FROM VERTICA my_database.' || schema_name || '.' || table_name || ' DIRECT;' from ALL_TABLES where schema_name not like 'v_%' and table_type='TABLE' order by schema_name);

SELECT 'CONNECT TO VERTICA my_database USER dbadmin PASSWORD ''XXX'' ON ''SOURCE_IP_ADDRESS'',5433;' union all (select 'select ''copying....' || schema_name || '.' || table_name || '''; ' || 'COPY ' || schema_name || '.' || table_name || ' FROM VERTICA my_database.' || schema_name || '.' || table_name || ' DIRECT;' from ALL_TABLES where schema_name not like 'v_%' and table_type='TABLE' order by schema_name);

That’s all there is to it. You’ve got yourself an Eon cluster!

这里的所有都是它的。 您已经拥有一个Eon集群！

My take away …

我带走了……

1. Compute cost reduction to one-third.
   计算成本降低到三分之一。
2. Cost can be as closer to zero dollars since Eon allows you to scale down to even zero nodes when you don’t need to use your database. (off working hours) Your data is never permanently sitting on the database nodes. It resides in communal storage like an S3 bucket. You can simply bring it back online by cranking up a node whenever you wish to make your database available.
   成本可以接近零美元，因为当您不需要使用数据库时，Eon允许您缩小到甚至零个节点。 (下班时间)您的数据永远不会永久地位于数据库节点上。 它像S3存储桶一样位于公共存储中。 只要您希望使数据库可用，就可以通过启动一个节点来简单地使其重新联机。
3. Very low maintenance in comparison to Enterprise mode.
   与企业模式相比，维护成本非常低。
4. Less time spent when Vertica goes down (whatever the reason may be).
   Vertica发生故障时花费的时间更少(无论原因是什么)。
5. Create brand new Eon Vertica cluster quicker and bring it back up and running with data in current state as before.
   更快地创建全新的Eon Vertica集群，并像以前一样将其备份并以当前状态运行。
6. Very easy to switch or replicate your database to another location.
   将数据库切换或复制到另一个位置非常容易。

We can help you convert your Enterprise to Eon as well.

我们也可以帮助您将企业转换为Eon。

I recommend using www.elasticdw.com. You get all Eon/Enterprise related task done in pretty much no time and with zero maintenance on your side.

我建议使用www.elasticdw.com 。 您几乎可以立即完成所有与Eon / Enterprise相关的任务，并且只需零维护 。

翻译自: https://insight.full360.com/converting-enterprise-vertica-to-eon-431af58f9b70

vertica