Update the operating system

Needs dev input: remove before publishing

TODO: Netwrix doesn't pin a Docker version, and test labs upgrade without problems on Docker 29.x. Decide as a policy call (product management or Support) whether to keep the hold on the docker-ce, docker-ce-cli, and containerd.io packages in this procedure.

This page applies to the deprecated ISO appliance. For the supported deployment method, see Legacy ISO installation.

Install operating system updates during a scheduled maintenance window, because an update can require a reboot of the server or the Docker service. The downtime depends on the deployment:

Three-node cluster: No data loss. The API and background services restart briefly when a node reboots. Update one node at a time.
Single node: Downtime during the reboot, usually 30 to 60 minutes.

Prerequisites

A scheduled maintenance window. For a single node, plan for 30 to 60 minutes of downtime.
The s1 command-line helper, installed on the nodes.
Shell access with sudo rights to each node.

Update a three-node cluster

Update one node at a time.

On any node, open a database shell:
```
sudo s1 db
```
Check the replication status. Each member must show PRIMARY or SECONDARY in stateStr, and the replication lag must be a few seconds at most:
```
rs.status().members.forEach(function (m) { print(m.name, m.stateStr) })
```
Then check the replication lag. On MongoDB 4.0, run:
```
rs.printSlaveReplicationInfo()
```
On MongoDB 8.0, run:
```
rs.printSecondaryReplicationInfo()
```
If replication isn't healthy, resolve the problem before you continue.
Exit the database shell:
```
exit
```
On the primary node, check the status of the services and nodes:
```
s1 status; s1 nodes
```
Connect by SSH to the node that you update.
Check that the Docker packages have a hold:
```
sudo apt-mark showhold
```
If the command doesn't list docker-ce, docker-ce-cli, and containerd.io, add them to the hold list:
```
sudo apt-mark hold docker-ce docker-ce-cli containerd.io
```
Download the package information and list the available updates:
```
sudo apt update; sudo apt list --upgradable
```
Install the updates:
```
sudo apt -y upgrade
```
If a prompt asks whether to replace a configuration file or setting, keep the existing configuration, setting, or file.
Check whether a reboot is required:
```
sudo cat /var/run/reboot-required
```
If a reboot is required, the command returns *** System restart required ***. Otherwise, it returns cat: /var/run/reboot-required: No such file or directory.
If a reboot is required, reboot the node:
```
sudo reboot
```
From another node, wait until the rebooted node completes the restart:
```
watch s1 nodes
```
When Availability shows Active for the updated node, press Ctrl+C to stop the command.
Check the replication status again, as described in steps 1 through 3.
Check the status of the services and nodes:
```
s1 status; s1 nodes
```
Repeat steps 5 through 14 for the next node.
After you update all nodes, check the status of the services and nodes one more time:
```
s1 status; s1 nodes
```

Update a single node

Check the status of the services and nodes:
```
s1 status; s1 nodes
```
Check that the Docker packages have a hold:
```
sudo apt-mark showhold
```
If the command doesn't list docker-ce, docker-ce-cli, and containerd.io, add them to the hold list:
```
sudo apt-mark hold docker-ce docker-ce-cli containerd.io
```
Download the package information and list the available updates:
```
sudo apt update; sudo apt list --upgradable
```
Install the updates:
```
sudo apt -y upgrade
```
If a prompt asks whether to replace a configuration file or setting, keep the existing configuration, setting, or file.
Check whether a reboot is required:
```
sudo cat /var/run/reboot-required
```
If a reboot is required, the command returns *** System restart required ***. Otherwise, it returns cat: /var/run/reboot-required: No such file or directory.
If a reboot is required, reboot the node:
```
sudo reboot
```
After the reboot completes, sign in and check the status of the services and nodes:
```
s1 status; s1 nodes
```

Prerequisites​

Update a three-node cluster​

Update a single node​

Prerequisites

Update a three-node cluster

Update a single node