Netdata Agent: Backup restore documentation (#17006)
A document on how to backup and restore the Netdata agent for the public docs. This works in a test environment and I've used it once in production so please check it over as it may need ammendments. --------- Co-authored-by: Fotis Voutsas <fotis@netdata.cloud>
Luis Johnstone committed
Feb 28, 2024 at 16:02 UTC
57ffef0c92c479a0bb5e48c87c90c31d660d8049
1 file changed
+66
docs/maintenance/backup-restore.md
new
+66
@@ -0,0 +1,66 @@
1
+# Backing up a Netdata Agent
2
+
3
+## Introduction
4
+
5
+When preparing to backup a Netdata Agent it is worth considering that there are different kinds of data that you may wish to backup independently or all together:
6
+
7
+| Data type | Description | Location |
8
+|---------------------|------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------|
9
+| Agent configuration | Files controlling configuration of the Netdata Agent | [config directory](https://github.com/netdata/netdata/blob/master/docs/netdata-agent/configuration.md) |
10
+| Metrics | Database files | /var/cache/netdata |
11
+| Identity | Claim token, API key and some other files | /var/lib/netdata |
12
+
13
+
14
+## Scenarios
15
+
16
+### Backing up to restore data in case of a node failure
17
+
18
+In this standard scenario you are backing up your Netdata Agent in case of a node failure or data corruption so that the metrics and the configuration can be recovered. The purpose is not to backup/restore the application itself.
19
+
20
+1. Verify that the directory-paths in the table above contain the information you expect.
21
+
22
+ > **Note**
23
+ > The specific paths may vary depending upon installation method, Operating System and whether it is a Docker/Kubernetes deployment.
24
+
25
+2. It is recommended that you [stop the Netdata Agent](https://github.com/netdata/netdata/blob/master/docs/configure/start-stop-restart.md) when backing up the Metrics/database files.
26
+ Backing up the Agent configuration and Identity folders is straight-forward as they should not be changing very frequently.
27
+
28
+3. Using a backup tool such as `tar` you will need to run the backup as _root_ or as the _netdata_ user in order to access all the files in the directories.
29
+
30
+ ```
31
+ sudo tar -cvpzf netdata_backup.tar.gz /etc/netdata/ /var/cache/netdata /var/lib/netdata
32
+ ```
33
+
34
+ Stopping the Netdata agent is mostly required in order to back up the _database files_ of the Netdata Agent.
35
+
36
+ If you wish to minimize the gap in metrics caused by stopping the Netdata Agent, then you could have a backup job or script that uses the following sequence:
37
+
38
+- Backup the Agent configuration Identity directories
39
+- Stop the Netdata service
40
+- Backup up up the database files
41
+- Restart the netdata agent.
42
+
43
+### Restoring Netdata
44
+
45
+1. Ensure that the Netdata agent is installed and is [stopped](https://github.com/netdata/netdata/blob/master/packaging/installer/README.md#maintaining-a-netdata-agent-installation)
46
+
47
+ If you plan to deploy the Agent and restore a backup on top of it, then you might find it helpful to use the [`--dont-start-it`](https://github.com/netdata/netdata/blob/master/packaging/installer/methods/kickstart.md#other-options) option upon installation.
48
+
49
+ ```
50
+ wget -O /tmp/netdata-kickstart.sh https://get.netdata.cloud/kickstart.sh && sh /tmp/netdata-kickstart.sh --dont-start-it
51
+ ```
52
+
53
+ > **Note**
54
+ > If you are going to restore the database files then you should first ensure that the Metrics directory is empty.
55
+ >
56
+ > ```
57
+ > sudo rm -Rf /var/cache/netdata
58
+ > ```
59
+
60
+2. Restore the backup from the archive
61
+
62
+ ```
63
+ sudo tar -xvpzf /path/to/netdata_backup.tar.gz -C /
64
+ ```
65
+
66
+3. [Start the Netdata agent](https://github.com/netdata/netdata/blob/master/packaging/installer/README.md#maintaining-a-netdata-agent-installation)