# Cluster build log: Moving to temporary cluster

DevFeed: [Cluster build log: Moving to temporary cluster](<https://devfeed.tech/articles/cluster-build-log-moving-to-temporary-cluster-34848.md>)

Original publisher: [Read original article](<https://pine64.org/2021/09/01/clusters-build-log-moving-to-temporary-cluster/>)

Published: 2021-09-01T00:00:00Z

Content type: article

Language: en

Sources: [Community blog on PINE64](<https://devfeed.tech/sources/community-blog-on-pine64.md>)

Topics: [hosting](<https://devfeed.tech/topics/hosting.md>), [Network](<https://devfeed.tech/topics/network.md>), [Ethernet](<https://devfeed.tech/topics/ethernet.md>)

Tags: [cluster](<https://devfeed.tech/tags/cluster.md>), [community](<https://devfeed.tech/tags/community.md>), [downtime](<https://devfeed.tech/tags/downtime.md>), [ethernet](<https://devfeed.tech/tags/ethernet.md>), [hosting](<https://devfeed.tech/tags/hosting.md>), [network](<https://devfeed.tech/tags/network.md>), [rockpro64-2](<https://devfeed.tech/tags/rockpro64-2.md>)

## AI overview

This build log explains why PINE64 is creating a temporary hosting cluster. The existing cluster needs hardware changes that require extended downtime, so a smaller cluster will keep community services running during the work. The article introduces the power and network requirements for the temporary cluster.

## Source excerpt

As promised in the last month's community update Housekeeping section, this is the first part of the cluster's build log! This mini-series in which I will cover everything about upcoming changes our hosting cluster, and it begins with building a temporary cluster. Why build a temporary cluster? One of the latest photo of cluster Let's start from beginning. When we built the current hosting cluster a year ago, it was made in quite a hurry (some of the services needed to be migrated to the cluster on a moment's notice). Also, it was the first time I (gamiee) and Matthew (fire219) were tasked with building a cluster. Because of this, some of the cluster's features are missing and various things don't work quite as well as we have hoped. For example, the cluster has power management boards (PCB on right side of every row), which are used primary for power distribution to the SBCs, but also it also has a I2C GPIO Expander for rebooting and turning off the SBCs. The I2C Expansion isn't in use yet, because the primary node wasn't able to detect it, so this is something that needs to be checked out when the cluster is turned off. And this is exactly the main reason why we are building a temporary cluster. We need to do some changes which require us to have the whole cluster turned off and on a workbench (not in housing room where isn't much space), and this means that it will be inaccessible for several days. Obviously a long downtime wouldn't be good when 100% of PINE64 community services are running on it. So we are going to build a temporary small cluster, which will have the minimal number of required nodes to have all services running, so we can make all upgrades with precision and at our own pace. What do we need to make a custom cluster? There are two things each node needs: a power supply and access to the network. Access to the network can be solved by using a classic Ethernet switch, but powering multiple RockPros, which are actually pretty power hungry, is a bit