# Build doesn't finish after completing tests

**URL:** <https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288>\
**Category:** Windows\
**Created:** [October 12, 2018, 12:32pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288 "2018-10-12T12:32:13Z")\
**Posts on this page:** 20\
**Page:** 1

<div class="post-metadata">

**Author:** ![xPaw](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/xpaw/32/148_2.png) [@xPaw](https://travis-ci.community/u/xPaw)\
**Post date:** [October 12, 2018, 12:32pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/1 "2018-10-12T12:32:13Z")

</div>

Looks like this:

> Done. Your build exited with 0.  
> No output has been received in the last 10m0s, this potentially indicates a stalled build or something wrong with the build itself.

> **[Travis CI - Test and Deploy with Confidence](https://app.travis-ci.com/thelounge/thelounge/jobs/440606641)**
>
> Travis CI enables your team to test and ship your apps with confidence. Easily sync your projects with Travis CI and you'll be testing your code in minutes.

> **[Travis CI - Test and Deploy with Confidence](https://app.travis-ci.com/thelounge/thelounge/jobs/440339972)**
>
> Travis CI enables your team to test and ship your apps with confidence. Easily sync your projects with Travis CI and you'll be testing your code in minutes.

This doesn’t happen on linux or osx builds.

---

<div class="post-metadata">

**Author:** ![BanzaiMan](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/banzaiman/32/67_2.png) [@BanzaiMan](https://travis-ci.community/u/BanzaiMan)\
**Post date:** [October 12, 2018, 12:46pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/2 "2018-10-12T12:46:03Z")

</div>

Do you start some sort of subprocess in the background during your tests? If so, Windows might need extra nudge cleaning up these processes in order to exit the build cleanly.

---

<div class="post-metadata">

**Author:** ![xPaw](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/xpaw/32/148_2.png) [@xPaw](https://travis-ci.community/u/xPaw)\
**Post date:** [October 12, 2018, 1:31pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/3 "2018-10-12T13:31:43Z")

</div>

I don’t think we do.

I also develop on Windows myself, and running tests exists cleanly.

---

<div class="post-metadata">

**Author:** ![AviVahl](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/avivahl/32/399_2.png) [@AviVahl](https://travis-ci.community/u/AviVahl)\
**Post date:** [October 12, 2018, 1:52pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/4 "2018-10-12T13:52:38Z")

</div>

same issue here:  
[https://travis-ci.com/AviVahl/ts-tools/builds/87770857](https://travis-ci.com/AviVahl/ts-tools/builds/87770857)

using yarn & lerna. works on linux/osx. windows is stuck.  
might also be worth noting same repo works on appveyor.

thx for the great feature! would love to only use a single CI.

---

<div class="post-metadata">

**Author:** ![BanzaiMan](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/banzaiman/32/67_2.png) [@BanzaiMan](https://travis-ci.community/u/BanzaiMan)\
**Post date:** [October 12, 2018, 5:33pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/5 "2018-10-12T17:33:31Z")

</div>

@AviVahl Do you start `gpg-tools`? This is what I see when I just add `ps -ef` after your build ends:

[https://staging.travis-ci.com/BanzaiMan/ts-tools/jobs/232474#L254](https://staging.travis-ci.com/BanzaiMan/ts-tools/jobs/232474#L254)

We should probably clean up child processes of the build script after the build for better UX.

---

<div class="post-metadata">

**Author:** ![xPaw](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/xpaw/32/148_2.png) [@xPaw](https://travis-ci.community/u/xPaw)\
**Post date:** [October 12, 2018, 7:03pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/6 "2018-10-12T19:03:10Z")

</div>

I just did `ps -ef` on our build, and we have `gpg-agent` in there too. I am quite certain that we don’t use it though.

[https://travis-ci.org/thelounge/thelounge/jobs/440769451#L221](https://travis-ci.org/thelounge/thelounge/jobs/440769451#L221)

Does it automatically start on the build machine (perhaps after running yarn?)

EDIT: This post was marked as spam?

---

<div class="post-metadata">

**Author:** ![AviVahl](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/avivahl/32/399_2.png) [@AviVahl](https://travis-ci.community/u/AviVahl)\
**Post date:** [October 12, 2018, 9:56pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/7 "2018-10-12T21:56:52Z")

</div>

My guess is the yarn installation:

```auto
before_install:
  - curl -o- -L https://yarnpkg.com/install.sh | bash -s -- --version 1.10.1
  - export PATH=$HOME/.yarn/bin:$PATH

```

their install.sh uses gpg.

The above config is the one suggested by the Travis docs for custom yarn version.

---

<div class="post-metadata">

**Author:** ![AviVahl](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/avivahl/32/399_2.png) [@AviVahl](https://travis-ci.community/u/AviVahl)\
**Post date:** [October 12, 2018, 10:13pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/8 "2018-10-12T22:13:05Z")

</div>

Checked it without my `before_install`. Still got stuck. It appears the yarn installation (and pgp verification) occurs anyway, because my repo contains a `yarn.lock`.

The repo is setup using yarn workspaces, so I have no proper matching alternative in npm land yet. ☹

EDIT:  
tried setting env variable `YARN_GPG=no` (via travis’s configuration) to force no gpg validation, but the container appears to be stuck on: `Worker information`

---

<div class="post-metadata">

**Author:** ![AviVahl](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/avivahl/32/399_2.png) [@AviVahl](https://travis-ci.community/u/AviVahl)\
**Post date:** [October 13, 2018, 11:12am UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/9 "2018-10-13T11:12:56Z")

</div>

> [@AviVahl](#):
>
> YARN\_GPG=no

I’ve used secure env tokens (via Travis config- “hide from log”), which seems to be broken with Windows containers. Trying regular env var to see if behavior is different.

EDIT:  
Yes! It was the gpg that install.sh of yarn runs. Build now passes, after setting the YARN\_GPG=no environment variable.

---

<div class="post-metadata">

**Author:** ![kris](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/kris/32/13_2.png) [@kris](https://travis-ci.community/u/kris)\
**Post date:** [October 13, 2018, 9:33pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/10 "2018-10-13T21:33:19Z")

</div>

> [@xPaw](#):
>
> EDIT: This post was marked as spam?

Sorry about that - I think I’ve fixed it, but please let us know if it happens again. Thanks!

---

<div class="post-metadata">

**Author:** ![chrisdothtml](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/chrisdothtml/32/224_2.png) [@chrisdothtml](https://travis-ci.community/u/chrisdothtml)\
**Post date:** [October 18, 2018, 1:45pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/11 "2018-10-18T13:45:46Z")

</div>

Similar problem here, but my builds are hanging after failing

[https://travis-ci.org/chrisdothtml/hey/builds/443190221](https://travis-ci.org/chrisdothtml/hey/builds/443190221)

---

<div class="post-metadata">

**Author:** ![josh](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/josh/32/12_2.png) [@josh](https://travis-ci.community/u/josh)\
**Post date:** [October 20, 2018, 11:45pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/13 "2018-10-20T23:45:13Z")

</div>

Hi @chrisdothtml

Could you please try adding `YARN_GPG=no` to your `.travis.yml` config to see if this helps?

@BanzaiMan do we currently detect and kill rogue processes when a build finishes?

---

<div class="post-metadata">

**Author:** ![chrisdothtml](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/chrisdothtml/32/224_2.png) [@chrisdothtml](https://travis-ci.community/u/chrisdothtml)\
**Post date:** [October 21, 2018, 1:38am UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/14 "2018-10-21T01:38:28Z")

</div>

@josh Tried that. New error this time

[https://travis-ci.org/chrisdothtml/hey/jobs/444241108](https://travis-ci.org/chrisdothtml/hey/jobs/444241108)

---

<div class="post-metadata">

**Author:** ![josh](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/josh/32/12_2.png) [@josh](https://travis-ci.community/u/josh)\
**Post date:** [October 21, 2018, 1:43am UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/15 "2018-10-21T01:43:35Z")

</div>

@chrisdothtml I’ve seen this time to time, can you restart to see if the issue persists?

---

<div class="post-metadata">

**Author:** ![chrisdothtml](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/chrisdothtml/32/224_2.png) [@chrisdothtml](https://travis-ci.community/u/chrisdothtml)\
**Post date:** [October 21, 2018, 1:55am UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/16 "2018-10-21T01:55:24Z")

</div>

@josh yep, looks like restarting the build resolved it. Thanks!

---

<div class="post-metadata">

**Author:** ![josh](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/josh/32/12_2.png) [@josh](https://travis-ci.community/u/josh)\
**Post date:** [October 21, 2018, 1:58am UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/17 "2018-10-21T01:58:21Z")

</div>

Fantastic to hear!

I will add this issue to one we should address long term.

---

<div class="post-metadata">

**Author:** ![s-light](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/s-light/32/2002_2.png) [@s-light](https://travis-ci.community/u/s-light)\
**Post date:** [June 18, 2019, 10:27pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/18 "2019-06-18T22:27:03Z")

</div>

@josh i encountered the original ‘all fine but dont finishing’ today. and also the fix with `YARN_GPG=no` worked.  
is there a bug that i can track?

and by the way -  
is there any option to set YARN\_GPG only on windows builds?

sunny greetings  
stefan

---

<div class="post-metadata">

**Author:** ![DanielRuf](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/danielruf/32/2044_2.png) [@DanielRuf](https://travis-ci.community/u/DanielRuf)\
**Post date:** [June 21, 2019, 4:25pm UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/19 "2019-06-21T16:25:21Z")

</div>

This happens again. 2 different projects never finish the Windows builds.

---

<div class="post-metadata">

**Author:** ![DanySK](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/danysk/32/1399_2.png) [@DanySK](https://travis-ci.community/u/DanySK)\
**Post date:** [July 4, 2019, 8:04am UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/20 "2019-07-04T08:04:15Z")

</div>

Same issue here. Windows builds do not terminate. [https://travis-ci.org/DanySK/jirf/builds/554123269](https://travis-ci.org/DanySK/jirf/builds/554123269)

---

<div class="post-metadata">

**Author:** ![DanySK](https://sea1.discourse-cdn.com/flex015/user_avatar/travis-ci.community/danysk/32/1399_2.png) [@DanySK](https://travis-ci.community/u/DanySK)\
**Post date:** [July 4, 2019, 8:50am UTC](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288/21 "2019-07-04T08:50:15Z")

</div>

OK, in my case was the Gradle daemon hanging the build. I was able to configure the build for Java / Gradle via Jabba. It’s been kind of a hell though.

[Next page](https://travis-ci.community/t/build-doesnt-finish-after-completing-tests/288.md?page=2)
