kubectl-drain(1)

1KUBERNETES(1)(kubernetes)                            KUBERNETES(1)(kubernetes)
2
3
4
5Eric Paris Jan 2015
6
7

NAME

9       kubectl drain - Drain node in preparation for maintenance
10
11
12

SYNOPSIS

14       kubectl drain [OPTIONS]
15
16
17

DESCRIPTION

19       Drain node in preparation for maintenance.
20
21
22       The  given  node  will be marked unschedulable to prevent new pods from
23       arriving.  'drain'  evicts  the  pods  if  the  API   server   supports
24       https://kubernetes.io/docs/concepts/workloads/pods/disruptions/   evic‐
25       tion https://kubernetes.io/docs/concepts/workloads/pods/disruptions/  .
26       Otherwise,  it  will  use normal DELETE to delete the pods. The 'drain'
27       evicts or deletes all pods except mirror pods (which cannot be  deleted
28       through  the  API server).  If there are daemon set-managed pods, drain
29       will not proceed without --ignore-daemonsets, and  regardless  it  will
30       not delete any daemon set-managed pods, because those pods would be im‐
31       mediately replaced by the daemon  set  controller,  which  ignores  un‐
32       schedulable  markings.   If  there are any pods that are neither mirror
33       pods nor managed by a replication controller, replica set, daemon  set,
34       stateful  set,  or  job, then drain will not delete any pods unless you
35       use --force.  --force will also allow deletion to proceed if the manag‐
36       ing resource of one or more pods is missing.
37
38
39       'drain'  waits  for graceful termination. You should not operate on the
40       machine until the command completes.
41
42
43       When you are ready to put the node back into service, use  kubectl  un‐
44       cordon, which will make the node schedulable again.
45
46
47       https://kubernetes.io/images/docs/kubectl_drain.svg Workflowhttps://ku‐
48       bernetes.io/images/docs/kubectl_drain.svg
49
50
51

OPTIONS

53       --chunk-size=500      Return large lists in chunks rather than  all  at
54       once.  Pass  0  to disable. This flag is beta and may change in the fu‐
55       ture.
56
57
58       --delete-emptydir-data=false      Continue even if there are pods using
59       emptyDir (local data that will be deleted when the node is drained).
60
61
62       --delete-local-data=false       Continue  even  if there are pods using
63       emptyDir (local data that will be deleted when the node is drained).
64
65
66       --disable-eviction=false      Force drain to use delete, even if  evic‐
67       tion  is supported. This will bypass checking PodDisruptionBudgets, use
68       with caution.
69
70
71       --dry-run="none"      Must be "none", "server", or "client". If  client
72       strategy, only print the object that would be sent, without sending it.
73       If server strategy, submit server-side request without  persisting  the
74       resource.
75
76
77       --force=false       Continue even if there are pods that do not declare
78       a controller.
79
80
81       --grace-period=-1      Period of time in seconds given to each  pod  to
82       terminate  gracefully.  If negative, the default value specified in the
83       pod will be used.
84
85
86       --ignore-daemonsets=false      Ignore DaemonSet-managed pods.
87
88
89       --pod-selector=""      Label selector to filter pods on the node
90
91
92       -l, --selector=""      Selector (label query) to  filter  on,  supports
93       '=', '==', and '!='.(e.g. -l key1=value1,key2=value2). Matching objects
94       must satisfy all of the specified label constraints.
95
96
97       --skip-wait-for-delete-timeout=0      If  pod  DeletionTimestamp  older
98       than N seconds, skip waiting for the pod.  Seconds must be greater than
99       0 to skip.
100
101
102       --timeout=0s      The length of time to wait  before  giving  up,  zero
103       means infinite
104
105
106

OPTIONS INHERITED FROM PARENT COMMANDS

108       --as=""      Username to impersonate for the operation. User could be a
109       regular user or a service account in a namespace.
110
111
112       --as-group=[]      Group to impersonate for the  operation,  this  flag
113       can be repeated to specify multiple groups.
114
115
116       --as-uid=""      UID to impersonate for the operation.
117
118
119       --azure-container-registry-config=""       Path  to the file containing
120       Azure container registry configuration information.
121
122
123       --cache-dir="/builddir/.kube/cache"      Default cache directory
124
125
126       --certificate-authority=""      Path to a cert file for the certificate
127       authority
128
129
130       --client-certificate=""      Path to a client certificate file for TLS
131
132
133       --client-key=""      Path to a client key file for TLS
134
135
136       --cluster=""      The name of the kubeconfig cluster to use
137
138
139       --context=""      The name of the kubeconfig context to use
140
141
142       --disable-compression=false       If true, opt-out of response compres‐
143       sion for all requests to the server
144
145
146       --insecure-skip-tls-verify=false      If true, the server's certificate
147       will not be checked for validity. This will make your HTTPS connections
148       insecure
149
150
151       --kubeconfig=""      Path to the kubeconfig file to  use  for  CLI  re‐
152       quests.
153
154
155       --match-server-version=false        Require  server  version  to  match
156       client version
157
158
159       -n, --namespace=""      If present, the namespace scope  for  this  CLI
160       request
161
162
163       --password=""      Password for basic authentication to the API server
164
165
166       --profile="none"         Name   of   profile   to   capture.   One   of
167       (none|cpu|heap|goroutine|threadcreate|block|mutex)
168
169
170       --profile-output="profile.pprof"      Name of the  file  to  write  the
171       profile to
172
173
174       --request-timeout="0"       The length of time to wait before giving up
175       on a single server request. Non-zero values  should  contain  a  corre‐
176       sponding time unit (e.g. 1s, 2m, 3h). A value of zero means don't time‐
177       out requests.
178
179
180       -s, --server=""      The address and port of the Kubernetes API server
181
182
183       --tls-server-name=""      Server name to  use  for  server  certificate
184       validation.  If  it  is  not provided, the hostname used to contact the
185       server is used
186
187
188       --token=""      Bearer token for authentication to the API server
189
190
191       --user=""      The name of the kubeconfig user to use
192
193
194       --username=""      Username for basic authentication to the API server
195
196
197       --version=false      Print version information and quit
198
199
200       --warnings-as-errors=false      Treat warnings received from the server
201       as errors and exit with a non-zero exit code
202
203
204

EXAMPLE

206                # Drain node "foo", even if there are pods not managed by a replication controller, replica set, job, daemon set or stateful set on it
207                kubectl drain foo --force
208
209                # As above, but abort if there are pods not managed by a replication controller, replica set, job, daemon set or stateful set, and use a grace period of 15 minutes
210                kubectl drain foo --grace-period=900
211
212
213
214

HISTORY

221       January  2015,  Originally compiled by Eric Paris (eparis at redhat dot
222       com) based on the kubernetes source material, but hopefully  they  have
223       been automatically generated since!
224
225
226
227Manuals                              User            KUBERNETES(1)(kubernetes)