# Restart entire cluster

**URL:** https://community.cratedb.com/t/restart-entire-cluster/1746
**Category:** CrateDB
**Tags:** getting-started
**Created:** [April 2, 2024, 2:37am UTC](https://community.cratedb.com/t/restart-entire-cluster/1746 "2024-04-02T02:37:23Z")
**Posts on this page:** 2
**Page:** 1

<div class="post-metadata">

### Author: ![role](https://sea2.discourse-cdn.com/flex020/user_avatar/community.cratedb.com/role/32/1216_2.png) [@role](https://community.cratedb.com/u/role)
#### Post date: [April 2, 2024, 2:37am UTC](https://community.cratedb.com/t/restart-entire-cluster/1746/1 "2024-04-02T02:37:23Z")

</div>

hi

I didn’t know English so I used a translator, but thank you for your reference. ㅠㅠ

It is currently being used with 4 master nodes and 20 data nodes.

When usage on an unspecified data node is over 90%

It can be seen that the usage of other unspecified data nodes is less than 50%.

Is it a clustering problem?

For this reason, a full reboot method is used.

In this case, data loss inevitably occurs.

cluster.graceful\_stop.min\_availability = full cannot be added because there is an issue of adding backup equipment.

Options for smooth clustering

Is there a way to store shards without adding additional equipment?

Please  
If you have any ideas please give the newbie grace

---

<div class="post-metadata">

### Author: ![hernanc](https://sea2.discourse-cdn.com/flex020/user_avatar/community.cratedb.com/hernanc/32/676_2.png) [@hernanc](https://community.cratedb.com/u/hernanc)
#### Post date: [April 2, 2024, 3:47pm UTC](https://community.cratedb.com/t/restart-entire-cluster/1746/2 "2024-04-02T15:47:57Z")

</div>

Hi,

> When usage on an unspecified data node is over 90%  
> It can be seen that the usage of other unspecified data nodes is less than 50%.

Is this about disk space usage or CPU or memory usage?

For disk space usage you could check the [Shards browser — CrateDB: Admin UI](https://cratedb.com/docs/crate/admin-ui/en/latest/shards.html) to see if rebalancing is happening or if maybe one node got larger shards because of table/partition properties.

For CPU/memory usage you may want to check if requests coming from the clients are getting balanced across the nodes.

These are all situations that can be addressed without a full cluster restart, if you want to share more details we would be happy to help.