Re: [kubernetes/kubernetes] Kubectl cp gives "tar: removing leading '/' from member names" warning (#58692)
Erick Fejta
@kubernetes/sig-cli-feature-requests
—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
Jesse Almanrode
Something I found is that if I do not put a / at the beginning of my path following : in <pod>:<path> that the kubectl cp command works fine.
Stephan Westen
I tried your suggestion to drop the leading slash like this:
kubectl cp cms-sw-one-6bf986b657-65vbn:home/cms/tomcat/logs ~/documents -c cms --namespace=od'
This gives the following error:
stderr: tar: home/cms/tomcat/logs: Cannot stat: No such file or directory
With the leading / it works fine except for the warning.
Konstantin Chugalinskiy
You may try to copy your files to workdir and then retry to copy them using just their names. It's weird, but it works for now.
Yonatan Bitton
I have the same problem, to solve this, I used the advice @kchugalinskiy gave.
I moved the wanted file to the working dir in the pod (the directory which is automatically opened, when you open bash on it) -
user@podname:/usr/src# ls data.txt data.txt
In my case it is - /usr/src folder.
Then in my local bash terminal -
user@local:~$ kubectl cp podname:data.txt data.txt user@local:~$ ls data.txt data.txt
Matt McKellar-Spence
I'm receiving the same error message, but also finding the file copied correctly from the container to my local machine.
- Client: v1.10.1
- Server: v1.9.4-gke.1
- Platform: GKE
I'm not sure where to add the tar -C flag when using kubectl cp.
Suraj Deshmukh
$ ll total 0 $ kubectl cp web-6fc578559d-sns6p:/usr/share/httpd/noindex/index.html ./new tar: Removing leading `/' from member names $ echo $? 0 $ cat new | head -10 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd"><html><head> <meta http-equiv="content-type" content="text/html; charset=UTF-8"> <title>Apache HTTP Server Test Page powered by CentOS</title> <meta http-equiv="Content-Type" content="text/html; charset=UTF-8"> <!-- Bootstrap --> <link href="/noindex/css/bootstrap.min.css" rel="stylesheet"> <link rel="stylesheet" href="noindex/css/open-sans.css" type="text/css" /> <style type="text/css"><!--
it's not an error some condition since the exit code is 0. Also the functionality works fine, it's just the error which comes which is kinda confusing.
And this is anwered in multiple places on stackoverflow, https://stackoverflow.com/a/25929424 and https://unix.stackexchange.com/a/59244
TonyAdo
I would think this is not a bug, it's just what we expect and should not break any functionality.
AnthonyWC
It works fine but the message is just confusing.
fejta-bot
Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.
If this issue is safe to close now please do so with /close.
Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale
fejta-bot
Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.
If this issue is safe to close now please do so with /close.
Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten
fejta-bot
Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.
Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close
Kubernetes Prow Robot
@fejta-bot: Closing this issue.
In response to this:
Rotten issues close after 30d of inactivity.
Reopen the issue with/reopen.
Mark the issue as fresh with/remove-lifecycle rotten.Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.
Kubernetes Prow Robot
Closed #58692.
Tuan Anh Tran
i cannot copy the file anywhere else. permission issue.
is there another workaround?
Suraj Deshmukh
@tuananh can you please provide the steps you are performing and what errors do you see? Please elaborate on what you are trying to achieve?
Tuan Anh Tran
@surajssd im trying to get redis dump from container to my machine using this command
kubectl cp dev/redisdev-master-0:/bitnami/redis/data/dump.rdb ~/dump.rdb
I'm getting the above error
tar: Removing leading `/' from member names
Suraj Deshmukh
@tuananh was the file copied successfully if yes then this is not really an error. If not then there could be other issues. See #58692 (comment) for more info
Tuan Anh Tran
On Thu, Apr 4, 2019 at 2:30 PM Suraj Deshmukh <notifi...@github.com>
wrote:
> @tuananh <https://github.com/tuananh> was the file copied successfully if
> yes then this is not really an error. If not then there could be other
> issues. See #58692 (comment)
> for more info
>
> —
> You are receiving this because you were mentioned.
> Reply to this email directly, view it on GitHub
> or mute the thread
> <https://github.com/notifications/unsubscribe-auth/AAmSTp5rOGSyz1ZITvuuRI6Dg-zWYvjIks5vdaoGgaJpZM4Rppzx>
> .
ushell
The error reason is kubectl cp not use right pod file path.
example:
Pod WORKDIR set '/var/www/html'
if you want copy /var/www/html/foo.log to your location file folder
you will run kubectl cp POD_NAME:foo.log /Users/foo/foo.log , it work ok,
if run kubectl cp POD_NAME:/var/www/html/foo.log /Users/foo/foo.log will get warning.
Allen
When i in c: run
kubectl cp backend-68c49589fb-7dkkm:/fileStorage/20191231/cd28206c5e654390a815cc03bbc6d14e.png d:/xxx.jpg
error:
tar: Removing leading `/' from member names
remove the / after pod: like
kubectl cp backend-68c49589fb-7dkkm:fileStorage/20191231/cd28206c5e654390a815cc03bbc6d14e.png d:/xxx.jpg
it's ok.
—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub, or unsubscribe.
Eivind
@wangrqsh The issue has nothing to do with the drives on your local Windows computer. The latter works for you because the WORKDIR in your pod is /. Try setting a different WORKDIR in your pod, and you will also experience the warning.
balibebas
Saw the message trying to copy files out of a Pod. Didn't think to actually check the file system until I saw the suggestion from @AnthonyWC:
It works fine but the message is just confusing.
Looks like the message is just a warning that isn't squelched. Probably.
Daniel
this should probably be reopened
Kael
Erhhung Yuan
Why this issue should be closed solely because it's not a serious bug or error is baffling. It is obviously a very real and annoying issue of usability given how many people have complained about it. Asking people to copy/move the file into the WORKDIR first is not a solution—I could just as easily ignore the warning in the first place.
If an output message is confusing and serves no informational purpose to the user, then it should be suppressed by default—asking the operator to > /dev/null is not a solution, either.
Daniel
On Fri, Apr 2, 2021 at 3:52 PM Erhhung Yuan ***@***.***>
wrote:
> Why this issue should be closed solely because it's not a serious bug or
> usability* given how many people have complained about it. Asking people
> to copy/move the file into the WORKDIR first is not a solution—I could
> just as easily ignore the warning in the first place.
>
> to the user, then it should be suppressed by default—asking the operator to >
> /dev/null is not a solution, either.
>
> —
> Reply to this email directly, view it on GitHub
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/ABSSN2W7WVUY2JIOPYRBLCDTGZDCDANCNFSM4ENGTTYQ>
> .
>
--
Daniel Yakobian
Case Western Reserve University - Comp Sci
Shubham Joshi
Hi ! I want to pick this up. Can suppress the warning which is very concussing for the users.
Shubham Joshi
/reopen
Kubernetes Prow Robot
@joshi95: You can't reopen an issue/PR unless you authored it or you are a collaborator.
In response to this:
/reopen
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.
—
shemerk
I have this exact same issue.
what I found was the repeatedly executing the command eventually works (like 1 success out of 5-10 failures)
it's weird, probably has to due with the enveloping Linux environment, and to make matters worse: when the command DO work, it doesn't actually create the copied file anywhere....
I eventually gave up and simply copy-pasted the file's content
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are on a team that was mentioned.
luckychengww
you can use command like this to avoid "tar: removing leading '/' from member names)" error
kubectl exec -c ${containerPod} -n ${namespace} ${podName} -- sh -c "cd /data/ && tar cf - config.yml" | tar xf - -C /home/ubuntu/
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are on a team that was mentioned.
Emrah Becer
As a workaround, I piped the output to sed and removed this useless message.
| sed -r "s/tar: removing leading '/' from member names//"
I don't want to suppress all output. It can contain useful messages but I just don't want to see this one. Especially if you write a bash script, it floods your screen.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are on a team that was mentioned.