whentimeslows/website
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fixes in style and more on CDK (#5292)

Browse Source 
* Fixes in style and more on CDK

* Addressing review comments
This commit is contained in:
Konstantinos Tsakalozos
2017-10-03 03:35:36 +03:00
committed by Steve Perry
parent 2b57c44a26
commit e4fd3f81a8
12 changed files with 269 additions and 224 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/getting-started-guides/ubuntu/backups.md
+49 -48
View File
@@ -3,7 +3,11 @@ title: Backups
---
{% capture overview %}
This page shows you how to backup and restore data from the different deployed services in a given cluster.
The state of a Kubernetes cluster is kept in the etcd datastore.
This page shows how to backup and restore the etcd shipped with
the Canonical Distribution of Kubernetes. Backing up application specific data,
normally stored in a persistent volume, is outside the scope of this
document.
{% endcapture %}
{% capture prerequisites %}
@@ -11,19 +15,47 @@ This page assumes you have a working Juju deployed cluster.
{% endcapture %}
{% capture steps %}
## Exporting cluster data
## Snapshot etcd data
Exporting of cluster data is not supported at this time.
The `snapshot` action of the etcd charm allows the operator to snapshot
a running cluster's data for use in cloning,
backing up, or migrating to a new cluster.
## Restoring cluster data
juju run-action etcd/0 snapshot target=/mnt/etcd-backups
Importing of cluster data is not supported at this time.
- **param** target: destination directory to save the resulting snapshot archive.
## Exporting etcd data
Migrating etcd is a fairly easy task.
## Restore etcd data
Step 1: Snapshot your existing cluster. This is encapsulated in the `snapshot`
The etcd charm is capable of restoring its data from a cluster-data snapshot
via the `restore` action.
This comes with caveats and a very specific path to restore a cluster:
The cluster must be in a state of only having a single member. So it's best to
deploy a new cluster using the etcd charm, without adding any additional units.
```
juju deploy etcd new-etcd
```
The above code snippet will deploy a single unit of etcd, as 'new-etcd'
```
juju run-action etcd/0 restore target=/mnt/etcd-backups
```
Once the restore action has completed, evaluate the cluster health. If the unit
is healthy, you may resume scaling the application to meet your needs.
- **param** target: destination directory to save the existing data.
- **param** skip-backup: Don't backup any existing data.
## Migrating an etcd cluster
Using the above snapshot and restore operations, migrating etcd is a fairly easy task.
**Step 1:** Snapshot your existing cluster. This is encapsulated in the `snapshot`
action.
```
@@ -36,7 +68,7 @@ Results:
Action queued with id: b46d5d6f-5625-4320-8cda-b611c6ae580c
```
Step 2: Check the status of the action so you can grab the snapshot and verify
**Step 2:** Check the status of the action so you can grab the snapshot and verify
the sum. The copy.cmd result output is a copy/paste command for you to download
the exact snapshot that you just created.
@@ -68,56 +100,24 @@ juju scp etcd/0:/home/ubuntu/etcd-snapshots/etcd-snapshot-2016-11-09-02.41.47.ta
sha256sum etcd-snapshot-2016-11-09-02.41.47.tar.gz
```
Step 3: Deploy the new cluster leader, and attach the snapshot:
**Step 3:** Deploy the new cluster leader, and attach the snapshot:
```
juju deploy etcd new-etcd --resource snapshot=./etcd-snapshot-2016-11-09-02.41.47.tar.gz
```
Step 4: Re-Initialize the master with the data from the resource we just attached
**Step 4:** Reinitialize the master with the data from the resource we just attached
in step 3.
```
juju run-action new-etcd/0 restore
```
## Restoring etcd data
Allows the operator to restore the data from a cluster-data snapshot. This
comes with caveats and a very specific path to restore a cluster:
The cluster must be in a state of only having a single member. So it's best to
deploy a new cluster using the etcd charm, without adding any additional units.
```
juju deploy etcd new-etcd
```
> The above code snippet will deploy a single unit of etcd, as 'new-etcd'
```
juju run-action etcd/0 restore target=/mnt/etcd-backups
```
Once the restore action has completed, evaluate the cluster health. If the unit
is healthy, you may resume scaling the application to meet your needs.
- **param** target: destination directory to save the existing data.
- **param** skip-backup: Don't backup any existing data.
## Snapshot etcd data
Allows the operator to snapshot a running clusters data for use in cloning,
backing up, or migrating Etcd clusters.
juju run-action etcd/0 snapshot target=/mnt/etcd-backups
- **param** target: destination directory to save the resulting snapshot archive.
{% endcapture %}
{% capture discussion %}
# Known Limitations
## Known Limitations
#### Loss of PKI warning
@@ -125,10 +125,11 @@ If you destroy the leader - identified with the `*` text next to the unit number
all TLS pki will be lost. No PKI migration occurs outside
of the units requesting and registering the certificates.
> Important: Mismanaging this configuration will result in locking yourself
> out of the cluster, and can potentially break existing deployments in very
> strange ways relating to x509 validation of certificates, which affects both
> servers and clients.
**Caution:** Mismanaging this configuration will result in locking yourself
out of the cluster, and can potentially break existing deployments in very
strange ways relating to x509 validation of certificates, which affects both
servers and clients.
{: .caution}
#### Restoring from snapshot on a scaled cluster