[duplicate] odo watch deletes all files #3307
Assignees
Labels
flake
Categorizes issue or PR as related to a flaky test.
priority/High
Important issue; should be worked on before any other issues (except priority/Critical issue(s)).
Projects
Comments
openshift-ci-robot
added
the
kind/bug
|
It seems the issue is with pushing files steps I think it stuck in any one of these steps may be in the step Cluster load increases with consistently pushing and build files due to the multiple changes (file/dir create , delete and update in server.js file) in the context dir causing the issue. Just to confirm it i am going to keep the server.js file change in the first place for watch command. EDIT: It worked, i mean route successfully hit the application end point by keeping the server.js file change in the first place for watch command verification. |
|
To be split into two. @amitkrout to open another issue where we'll need dev inputs. |
amitkrout
added
the
priority/High
|
@amitkrout Can you please provide steps how it can be reproduced outside of our tests? |
@kadel So this output confirms that the issue is with watch command itself. Then what is the issue with the test script: Timeout is happening while curling for a particular text instead of watch command. This is happening because watch command fails to capture the actual failure due to faulty script https://github.com/openshift/odo/blob/master/tests/integration/devfile/utils/utils.go#L212. The fault lies in the go routine, actually we are not handling buffer output efficiently for each change in the source. Asper our current code it looks in the buffer for a string
|
|
Issue #3420 created for watch verification test script refactor |
amitkrout
changed the title
|
I tried the above scenario and everything works me Also the above script has some problems. |
|
Also the above script has some problems. mkdir {1...9} doesn't create folders named from 1 to 9, it creates a folder named {1...9}. yes, you are right but it does not matter as i am able to reproduce the issue with this script.
sed command flag works differently on macOS on which i tested. @mik-dass Its a flake, so try till reproduce it 😉 |
I have tried multiple times, still couldn't. Please provide the full verbose logs till the point of failure. |
|
It appears here again - #3462 (comment) |
@mik-dass Can you try by adding more create, delete and replace text command in the same shell script. Again i am saying its a flake. Do not loose hope, i am sure you will be able to reproduce it in few more try. |
If you can reproduce it, please paste the verbose watch output.
Please update the script and add more commands. |
I think it happens due to #3479 for devfile components. |
|
None of the attempts of mine have been successful so far. Push triggered by watch is never stuck for me after running the script. I have tried even by changing the sleep value in the script. |
I also spend some time to reproduce it. None of the time i hit the issue out of 6 try. @mik-dass The same issue is hit in one of my pr. You can see the verbose log here at - https://deck-ci.apps.ci.l2s4.p1.openshiftapps.com/view/gcs/origin-ci-test/pr-logs/pull/openshift_odo/3546/pull-ci-openshift-odo-master-v4.2-integration-e2e/1281939561458438144#1:build-log.txt%3A565 |
This is a different error than the one reported here. |
@mik-dass Yes, but see the end result. It is exactly the same what i was reproduced locally. When i reproduced it for first time, similar error i was getting while curling the url and searching for the updated string. The new string was not updated because push stuck in between as i mentioned above and also attached the screen shoot for reference. In CI we hit the same error but in a different way. may be what i reproduce locally is rare case but the error we hit in CI is the real flake that we have been trying to reproduce since months. So you can look into the recent failure CI steps in a followup issue. I will try to reproduce it again whenever i will get time. |
The new string was not updated maybe because of #3479.
This makes it a completely different issue then the current one. Also in the above screenshots or logs it is not clear at all. We should open a new issue with the logs. |
As i mentioned above I will try to reproduce it again with all the verbose log whenever i will get time. |
Please till then close this issue and open a new issue later on as the error you mentioned in the above logs is completely different. |
i can read it from here #3307 (comment) aswell.
So you need the same steps #3307 (comment) in another issue with verbose log ? |
Yes and the description of this issue is about
Yes. Or maybe just rename the title and description of the issue to match what's in the logs here #3307 (comment). |
|
Now i think i made a mistake by adding a similar failure for reference which is already filed. TBH now i am afraid of updating issue with reference that looks similar to it |
amitkrout
changed the title
This was referenced Sep 30, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
/kind bug
What versions of software are you using?
Operating System:
CI
Output of
odo version:master
How did you run odo exactly?
#3307 (comment)
Actual behavior
watch stuck
Expected behavior
watch should push the changes
Any logs, error output, etc?
The text was updated successfully, but these errors were encountered: