Manage Kubernetes clusters with Emacs.
Home Page: https://kubernetes-el.github.io/kubernetes-el
License: GNU General Public License v3.0
Manage your Kubernetes clusters with Emacs.
For full documentation, please visit kubernetes-el.github.io/kubernetes-el.
Disabling automatic refresh helped many to solve issue #100. For example one can use the following configuration:
(use-package kubernetes
:ensure t
:commands (kubernetes-overview)
:config
(setq kubernetes-poll-frequency 3600
kubernetes-redraw-frequency 3600))
When trying to load the overview, I get the following errors
Type v to switch between resources, and ? for usage.
pcase-memoize: equal first branch, yet different [2 times]
Error running timer ‘kubernetes-state-trigger-redraw’: (wrong-type-argument magit-section [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil])
pcase-memoize: equal first branch, yet different [3 times]
error in process sentinel: kubernetes-process-kill-quietly: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil]
error in process sentinel: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil]
pcase-memoize: equal first branch, yet different [21 times]
error in process sentinel: kubernetes-process-kill-quietly: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil]
error in process sentinel: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil]
pcase-memoize: equal first branch, yet different [3 times]
Error running timer ‘kubernetes-state-trigger-redraw’: (wrong-type-argument magit-section [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil])
error in process sentinel: kubernetes-process-kill-quietly: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil]
error in process sentinel: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil]
error in process sentinel: Quit [6 times]
pcase-memoize: equal first branch, yet different [3 times]
Error running timer ‘kubernetes-state-trigger-redraw’: (wrong-type-argument magit-section [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil])
error in process sentinel: kubernetes-process-kill-quietly: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil]
error in process sentinel: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil]
error in process sentinel: Quit [2 times]
I am on version 0.12.0 and have magit version 2.12.1
I get this error in overview:
Error running timer ‘kubernetes-state-trigger-redraw’: (wrong-type-argument arrayp nil) [3 times]
Remove unneeded elements before passing to emacs
Right now, the AST render function erases and rebuilds the whole buffer, which resets point. Investigate ways of preserving point between rendering runs.
One approach could be to preserve the line and column offset within the current section. More complicated would be to implement a render system that only changes parts of the buffer corresponding to changed parts of the AST.
Make the rendering logic functionally pure, then implement a suite of unit tests for the rendering.
Implement a basic config management popup like the one in master and get it working with the application state. Then, relaunch the go subprocess if the namespace is changed.
The contexts in my .kube/config do not have a namespace specified. This results in the following error when calling kubernetes-overview:
Debugger entered--Lisp error: (wrong-type-argument stringp nil)
propertize(nil face kubernetes-namespace)
kubernetes-contexts--render-current-context(((name . "minikube") (context (cluster . "minikube") (user . "team"))) nil)
kubectl uses the default namespace for these contexts. Perhaps kubernetes-el could use the same default?
If used against openshift you may hit the following error:
Error from server (Forbidden): User "iocanel" cannot list all namespaces in the cluster
which is a result of using:
kubectl get namespaces -o json --namespace=syndesis-ci
I think that for openshift it would make more sense to use something like:
oc get projects -o json --namespace=syndesis-ci
Buffers that provide live-refreshing features, e.g. the overview, are erased and re-populated with the current state on every redraw. Currently the position in the buffer is retained, but the window line position is lost. This is very noticeable when scrolling, as the window can shift unpredictably when a redraw occurs.
To resolve this, either save the window line during redraw (easy), or come up with a better way of updating the buffer (hard).
Implement interactive rollback of deployments from the overview buffer.
Steps to reproduce.
Assume that we have selected namespace1
Press 'v' and go to 'pods' view
Now press 'c' and change press 'n' to set different namespace2
Now the page is re-drawn with the default view with deployments.
I think we should retain the selected view while switching namespace.
If you hit D multiple times, multiple delete marks will be drawn.
Implement basic support for creating, replacing and applying resources from the overview.
For example, l l (that is, kubernetes-logs-fetch-all) asks the user to select a pod. But it should default the pod under point, like Magit does.
I was looking at a yaml array (container args) and noticed it wasn't outputting each element on a new line. Instead they're all smushed together on one line.
I'm not sure if this a bug in the yaml library or kubernetes-el.
Get deployments, much in the same way that pods are currently handled.
Relates to #34.
I don't have any idea what causes it. I'm using Emacs 25 from MacPorts, and I'm on revision 03afbbd5029339c6a4fa3fe4ff79b4818f062295 of kubernetes-mode.
$ emacs -nw
# wait for a few minutes on "kubernetes-overview".
Fatal error 6: Abort trap
Backtrace:
0 emacs 0x00000001000d67f5 emacs_backtrace + 325
1 emacs 0x00000001000b741d terminate_due_to_signal + 93
2 emacs 0x00000001000d46d3 emacs_abort + 19
3 emacs 0x00000001000d48a0 child_status_changed + 64
4 emacs 0x000000010018506a handle_child_signal + 234
5 emacs 0x00000001000d5d25 deliver_process_signal + 69
6 libsystem_platform.dylib 0x00007fff9f782b3a _sigtramp + 26
7 ??? 0x000000000000002f 0x0 + 47
8 emacs 0x000000010017d6c9 create_process + 1257
9 emacs 0x000000010017ca51 Fmake_process + 2705
10 emacs 0x00000001001395da Ffuncall + 938
11 emacs 0x0000000100177ac2 exec_byte_code + 2562
12 emacs 0x00000001001394f3 Ffuncall + 707
13 emacs 0x0000000100177ac2 exec_byte_code + 2562
14 emacs 0x00000001001394f3 Ffuncall + 707
15 emacs 0x0000000100177ac2 exec_byte_code + 2562
16 emacs 0x00000001001394f3 Ffuncall + 707
17 emacs 0x0000000100177ac2 exec_byte_code + 2562
18 emacs 0x00000001001394f3 Ffuncall + 707
19 emacs 0x000000010013993f Frun_hook_with_args + 255
20 emacs 0x00000001001397c4 Frun_hooks + 68
21 emacs 0x00000001001395da Ffuncall + 938
22 emacs 0x0000000100177ac2 exec_byte_code + 2562
23 emacs 0x00000001001394f3 Ffuncall + 707
24 emacs 0x00000001001395da Ffuncall + 938
25 emacs 0x0000000100177ac2 exec_byte_code + 2562
26 emacs 0x000000010013a55a funcall_lambda + 810
27 emacs 0x00000001001394f3 Ffuncall + 707
28 emacs 0x000000010013a00c call1 + 44
29 emacs 0x00000001000c3cc5 timer_check + 997
30 emacs 0x00000001000c2455 swallow_events + 133
31 emacs 0x0000000100180efa wait_reading_process_output + 3290
32 emacs 0x0000000100007e02 sit_for + 242
33 emacs 0x00000001000bfeb6 read_char + 5574
34 emacs 0x00000001000bcaf3 read_key_sequence + 2467
35 emacs 0x00000001000bb1e4 __command_loop_1_block_invoke + 596
36 emacs 0x00000001001ccf6f mac_autorelease_loop + 47
37 emacs 0x00000001000baea6 command_loop_1 + 790
38 emacs 0x0000000100137ed5 internal_condition_case + 261
39 emacs 0x00000001001ccf6f mac_autorelease_loop + 47
40 emacs 0x00000001000ca560 command_loop_2 + 16
...
^C^C^C^C^C^C^C^Z
[1] + 19509 suspended emacs -nw
$ emacs --version
GNU Emacs 25.2.2
Copyright (C) 2017 Free Software Foundation, Inc.
GNU Emacs comes with ABSOLUTELY NO WARRANTY.
You may redistribute copies of GNU Emacs
under the terms of the GNU General Public License.
For more information about these matters, see the file named COPYING.
$ uname -a
Darwin utkumovio.local 16.5.0 Darwin Kernel Version 16.5.0: Fri Mar 3 16:52:33 PST 2017; root:xnu-3789.51.2~3/RELEASE_X86_64 x86_64 i386 MacBookPro11,4 Darwin```
I'm running latest version of magit and kubernetes from melpa
Magit 20180313.1552, Git 2.13.3, Emacs 25.3.1, darwin
kubernetes-el: 0.12.0
kubectl commands run fine on the same machine.
Following is the output from Messages
Error running timer ‘kubernetes-state-trigger-redraw’: (wrong-type-argument magit-section [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil])
pcase-memoize: equal first branch, yet different [3 times]
kubernetes-ast-eval: Wrong type argument: magit-section, [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil ...] ...] ...]
pcase-memoize: equal first branch, yet different [3 times]
Error running timer ‘kubernetes-state-trigger-redraw’: (wrong-type-argument magit-section [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil])
pcase-memoize: equal first branch, yet different [3 times]
Error running timer ‘kubernetes-state-trigger-redraw’: (wrong-type-argument magit-section [eieio-class-tag--magit-section context nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in *kubernetes overview*> nil nil nil nil nil nil nil] nil] nil])
It would be nice to support running kubectl through tramp.
Use-case: Kubernetes is isolated from the internet. You need to ssh in to a cluster to get access.
Some of the faces provided by this package are copy-pasted from Magit. Search for builtin faces that make sense to use instead.
Currently, the kubectl exec popup ignores the configuration popup's KUBECONFIG setting.
This means that when execing into pods, we will fail if we need to use a non-default kubeconfig file.
Setting this to point to a file will only update the configurations window (which will work fine). When running the exec commands, we can see all the running pods, but the actual exec command will not run with the --kubeconfig option enabled.
I fix this in my own environment by adding
(push '"--kubeconfig=/path/to/my/kubecofig/kubeconfig.yaml" args)
at line 121 .... But there must be a better way!
Rocking package, by the way! I'm the envy of the office!
Hey,
When I press try to open the logs of a pod (l l) on the pods view (v pods <enter>), it is trying to run:
kubectl logs <container name> --namespace=
instead of:
kubectl --kubeconfig=<my config> logs <container name> --namespace=
Probably the same as #64 but for logging instead of exec
P.S. The namespace is blank there because it is trying to use the namespace set in config rather than the namespace of the selected pod and I have a workaround so I can see pods in all namespaces
The core implementation of the application state should be functionally pure. Probably something modelled after redux. This will enable easier unit testing of state transitions.
Magit draws a message line at the top magit-status buffers if an error occurs. This is would be nice to have.
Thanks for your work on kubernetes-el!
Calling kubernetes-use-context outside of kubernetes-overview causes the point in the current buffer to move to the start of the buffer. To reproduce:
M-x kubernetes-use-contextHi!
I'm trying to view the logs of a pod with two containers and it is giving me the same error you would get if you did kubectl logs mypod instead of kubectl logs mypod mycontainer
Error from server (BadRequest): the server rejected our request for an unknown reason (get pods mypod)
Process kubernetes-exec exited abnormally with code 1
is there a way of working around this?
Steps to reproduce:
M-x kubernetes-overview<type 'l' as in logs><enter pod name>ERROR!.Thanks!
It's now possible to build a Kubernetes client that learns of all possible resource types from the API, and also is able to access arbitrary resource types by name without a pre-compiled model for the types.
Would it be possible to do the same in kubernetes-el? I'd like to be able to list custom resources.
Emacs crashed after sometime connecting to the kubernetes cluster
Emacs on OSX is super slow at starting subprocesses, leading to performance issues on the current Emacs release. Emacs is also pretty slow at json parsing.
This could be resolved by implementing a custom client for the Kubernetes API that ships with this mode. Ideally, it would establish a long-lived connection to Kubernetes and push updates to Emacs.
Some nice-to-haves, if we want to replace kubectl completely
Would it be possible to add the ability to stream/read logs from all containers in a pod at once?
Agree on the metadata to be provided by the Go backend for basic functionality.
Relates to #34.
I installed the package from Melpa and after starting up minikube and running M-x kubernetes-overview I get:
Turning on magit-auto-revert-mode...done
Type v to switch between resources, and ? for usage.
error in process sentinel: kubernetes-process-kill-quietly: Wrong type argument: stringp, nil
error in process sentinel: Wrong type argument: stringp, nil
error in process sentinel: kubernetes-process-kill-quietly: Wrong type argument: stringp, nil
error in process sentinel: Wrong type argument: stringp, nil
My minikube version is v0.16.0
Emacs version is 25.1
Kubectl version is v1.5.2
I get following exception
Error running timer ‘kubernetes-state-trigger-redraw’: (wrong-type-argument magit-section [eieio-class-tag--magit-section context nil #<marker at 1 in kubernetes overview> nil nil nil nil nil [eieio-class-tag--magit-section context-container nil #<marker at 1 in kubernetes overview> nil nil nil nil nil [eieio-class-tag--magit-section root nil #<marker at 1 in kubernetes overview> nil nil nil nil nil nil nil] nil] nil])
Implement a nice way to start an ediff session for a local file and the corresponding live resource.
Error running timer ‘kubernetes-refresh’: (file-error "Creating pipe" "Too many open files") [3 times]
Error running timer ‘kubernetes-refresh’: (file-error "Opening pty" "Too many open files")
Error running timer ‘kubernetes-refresh’: (file-error "Creating pipe" "Too many open files") [22 times]
Error running timer ‘persistent-scratch-save’: (file-error "Opening output file" "Too many open files" "/Users/nislam/.emacs.d/data/.persistent-scratch.new")
Error running timer ‘kubernetes-refresh’: (file-error "Creating pipe" "Too many open files") [60 times]
Error running timer ‘persistent-scratch-save’: (file-error "Opening output file" "Too many open files" "/Users/nislam/.emacs.d/data/.persistent-scratch.new")
Error running timer ‘kubernetes-refresh’: (file-error "Creating pipe" "Too many open files") [36 times]
How can I specify my kubeconfig file?
Currently, the entire buffer must be redrawn to display a deletion mark, which introduces latency. It might also be the cause of a bug that makes the next line undeletable until point is manually moved.
Resolve by making updates on lines edit the buffer in place without triggering a redraw, in addition to updating the drawing state.
Turns out that since I only test with Evil, the keymaps I ship with this mode don't even work. 😅
It would be nice, if the overview buffer also displayed statefulsets along with deployments (if they exist).
Hi Chris -- I've been using kubernetes-el for a little while now, and it's awesome. Thank you!
OpenShift is a distribution of kubernetes that adds additional features that I would like to enable in kubernetes-el. My idea is to add an openshift-overview function that would just call kubernetes-overview, while enabling additional openshift functionality (exposed via the 'oc' command, which is a replacement for 'kubectl').
Are you amenable to accepting PRs for this?
Thanks again for this useful package!
Right now, the overview buffer is a dumping ground for various resources exposed by Kubernetes. It would be better to organise things more, like having pods being subsections of their deployments or something.
I'm going through the Kubernetes manual to make sure I understand how things can be logically structured before I finalise this.
Currently, you have to select a namespace to be shown. It would be nice to be able to see all namespaces.
It seems like this is partially implemented as you can just press enter when configuring a namespace (which doesn't set it to "default")
At runtime, we should test whether the go server exists, and download it if not.
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.