@@ -24,6 +24,11 @@ be sure to account for the number of developers you expect to use Coder, as well
24
24
as the resources they need to run their workspaces. See our guide on on
25
25
[ compute resources] ( ../../guides/admin/resources.md ) for additional information.
26
26
27
+ > In GKE version 1.24 and later, Docker-based [ node image types] ( https://cloud.google.com/kubernetes-engine/docs/concepts/node-images )
28
+ > are not supported. The examples below use ` ubuntu_containerd ` and
29
+ > ` cos_containerd ` to meet this requirement. Docker-based node images will
30
+ > prevent GKE cluster creation.
31
+
27
32
If you expect to provision GPUs to your Coder workspaces, you ** must** use a
28
33
general-purpose
29
34
[ N1 machine type] ( https://cloud.google.com/compute/docs/machine-types#gpus ) in
@@ -52,27 +57,30 @@ that makes the most sense for your location.
52
57
> See See [ Network Policies] ( ../requirements.md#network-policies ) for more
53
58
> information.
54
59
60
+ The sample scripts below create an ` e2-standard-4 ` instance with 2 nodes for
61
+ evaluation purposes, with a configuration to auto-scale to 8 nodes as more
62
+ developer workspace pods are created. Depending on your needs, you can choose
63
+ other sizes. See [ machine type
64
+ comparisons] ( https://cloud.google.com/compute/docs/machine-types#machine_type_comparison )
65
+ in particular [ general-purpose machine types like n1 and
66
+ e2] ( https://cloud.google.com/compute/docs/general-purpose-machines ) . See
67
+ [ requirements] ( ../requirements.md ) for help estimating your cluster size.
68
+
55
69
### Option 1: Cluster with full support of Coder features
56
70
57
71
This option uses an Ubuntu node image to enable support of
58
72
[ Container-based Virtual Machines (CVMs)] ( ../../admin/workspace-management/cvms.md ) ,
59
73
allowing system-level functionalities such as Docker in Docker.
60
74
61
- > Please note that the sample script creates a ` n1-highmem-4 ` instance;
62
- > depending on your needs, you can choose a
63
- > [ larger size] ( https://cloud.google.com/compute/docs/machine-types#machine_type_comparison )
64
- > instead. See [ requirements] ( ../requirements.md ) for help estimating your
65
- > cluster size.
66
-
67
75
``` console
68
76
gcloud beta container --project "$PROJECT_ID" \
69
77
clusters create "$NEW_CLUSTER_NAME" \
70
78
--zone "$ZONE" \
71
79
--no-enable-basic-auth \
72
80
--node-version "latest" \
73
81
--cluster-version "latest" \
74
- --machine-type "n1-highmem -4" \
75
- --image-type "UBUNTU " \
82
+ --machine-type "e2-standard -4" \
83
+ --image-type "ubuntu_containerd " \
76
84
--disk-type "pd-standard" \
77
85
--disk-size "50" \
78
86
--metadata disable-legacy-endpoints=true \
@@ -99,20 +107,14 @@ This option uses a Container-Optimized OS (COS) and meets Coder's minimum
99
107
requirements. It does _ not_ enable the use of
100
108
[ CVMs] ( ../../admin/workspace-management/cvms.md ) .
101
109
102
- > Please note that the sample script creates a ` n1-highmem-4 ` instance;
103
- > depending on your needs, you can choose a
104
- > [ larger size] ( https://cloud.google.com/compute/docs/machine-types#machine_type_comparison )
105
- > instead. See [ requirements] ( ../requirements.md ) for help estimating your
106
- > cluster size.
107
-
108
110
``` console
109
111
gcloud beta container --project "$PROJECT_ID" \
110
112
clusters create "$NEW_CLUSTER_NAME" \
111
113
--zone "$ZONE" \
112
114
--no-enable-basic-auth \
113
115
--cluster-version "latest" \
114
- --machine-type "n1-highmem -4" \
115
- --image-type "COS " \
116
+ --machine-type "e2-standard -4" \
117
+ --image-type "cos_containerd " \
116
118
--disk-type "pd-standard" \
117
119
--disk-size "50" \
118
120
--metadata disable-legacy-endpoints=true \
0 commit comments