kubernetes
/
kops
mirror of https://github.com/kubernetes/kops.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Cluster upgrade migration

This commit is contained in:
mikesplain 2019-09-14 11:23:36 -04:00
parent c818faedd2
commit 6294c96f01
4 changed files with 9 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/README.md
View File

@ -52,7 +52,7 @@
* [Cluster addons](addons.md)
* [Cluster configuration management](changing_configuration.md)
* [Cluster desired configuration creation from template](cluster_template.md)
* [Cluster upgrades and migrations](cluster_upgrades_and_migrations.md)
* [Cluster upgrades and migrations](operations/cluster_upgrades_and_migrations.md)
* [`etcd` volume encryption setup](etcd_volume_encryption.md)
* [`etcd` backup/restore](etcd/backup-restore.md)
* [GPU setup](gpu.md)

4
docs/cluster_upgrades_and_migrations.md → docs/operations/cluster_upgrades_and_migrations.md
View File

@ -5,7 +5,7 @@ At some point you will almost definitely want to upgrade the Kubernetes version
- Upgrade an existing `kube-up` managed cluster to one managed by `kops`
+ [The simple method with downtime](#kube-up---kops-downtime)
+ [The more complex method with zero-downtime](#kube-up---kops-sans-downtime)
- [Upgrade a `kops` cluster from one Kubernetes version to another](upgrade.md)
- [Upgrade a `kops` cluster from one Kubernetes version to another](../upgrade.md)
## `kube-up` -> `kops`, with downtime
@ -188,7 +188,7 @@ This method provides zero-downtime when migrating a cluster from `kube-up` to `k
Limitations:
- If you're using the default networking (`kubenet`), there is a account limit of 50 entries in a VPC's route table. If your cluster contains more than ~25 nodes, this strategy, as-is, will not work.
+ Shifting to a CNI-compatible overlay network like `weave`, `kopeio-vxlan` (`kopeio`), `calico`, `canal`, `romana`, and similar. See the [kops networking docs](networking.md) for more information.
+ Shifting to a CNI-compatible overlay network like `weave`, `kopeio-vxlan` (`kopeio`), `calico`, `canal`, `romana`, and similar. See the [kops networking docs](../networking.md) for more information.
+ One solution is to gradually shift traffic from one cluster to the other, scaling down the number of nodes on the old cluster, and scaling up the number of nodes on the new cluster.
### Steps

2
docs/upgrade.md
View File

@ -6,7 +6,7 @@ The `kops upgrade` command also automates checking for and applying updates.
It is recommended to run the latest version of Kops to ensure compatibility with the target kubernetesVersion. When applying a Kubernetes minor version upgrade (e.g. `v1.5.3` to `v1.6.0`), you should confirm that the target kubernetesVersion is compatible with the [current Kops release](https://github.com/kubernetes/kops/releases).
Note: if you want to upgrade from a `kube-up` installation, please see the instructions for [how to upgrade kubernetes installed with kube-up](cluster_upgrades_and_migrations.md).
Note: if you want to upgrade from a `kube-up` installation, please see the instructions for [how to upgrade kubernetes installed with kube-up](operations/cluster_upgrades_and_migrations.md).
### Manual update

2
mkdocs.yml
View File

@ -71,7 +71,7 @@ nav:
- Cluster Addons & Manager : "operations/addons.md"
- Cluster configuration management: "changing_configuration.md"
- Cluster Templating: "operations/cluster_template.md"
- Cluster upgrades and migrations: "cluster_upgrades_and_migrations.md"
- Cluster upgrades and migrations: "operations/cluster_upgrades_and_migrations.md"
- etcd volume encryption setup: "etcd_volume_encryption.md"
- etcd backup/restore: "etcd/backup-restore.md"
- GPU setup: "gpu.md"