Travis test gives error sometimes which gets removed by rebuild #2824
Assignees
Labels
help wanted
requires help by anyone willing to contribute
Ruby
testing
issues are usually for adding unit tests, integration tests or any other tests for a feature
Comments
Souravirus
added
testing
This was referenced Jun 14, 2018
Closed
|
This is causing a lot of trouble!! It fails often. @icarito @publiclab/reviewers if you have any ideas, or could do some research, that'd be great! |
|
Maybe the Debian Stretch update will help... #2865 |
|
Hmm, i just saw this on https://travis-ci.org/publiclab/plots2/builds/395173961 so I think it wasn't resolved... too bad! |
|
I usually get the following error which is resolved by rebuilding Travis
|
|
This might help |
|
OK, captured a log from this: https://gist.github.com/jywarren/6e06ae3a6647fadcf90dc7842416c5b9 |
This was referenced Jul 19, 2018
|
Saw another memory-related error in https://travis-ci.org/publiclab/plots2/builds/405981763 Restarting it... |
|
ok lets see about it. |
|
OK, so one plan is to try some of these Travis options. Another is to try running these in CircleCI or another service to see! |
|
Ok reading these options. Thanks |
|
Can I work on this after working on #2659? |
|
Sure, and I think @icarito is trying a few things out too!
…On Fri, Jul 27, 2018 at 5:57 PM Sourav Sahoo ***@***.***> wrote:
Can I work on this after working on #2659
<#2659>?
—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub
<#2824 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AABfJ_MDuQE3u7QUB6nJIjAYvc0P4nDIks5uK4yjgaJpZM4UjKqu>
.
|
|
thank you!
…On Fri, Jul 27, 2018 at 5:57 PM Sourav Sahoo ***@***.***> wrote:
Can I work on this after working on #2659
<#2659>?
—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub
<#2824 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AABfJ_MDuQE3u7QUB6nJIjAYvc0P4nDIks5uK4yjgaJpZM4UjKqu>
.
|
|
Hi! https://gitlab.com/publiclab/plots2/-/jobs/86207162 Has anyone found this error during local development? Thanks for your input! |
|
Hi, |
|
Okay I've captured a core dump and got a backtrace from gdb! https://gist.github.com/jywarren/6265fcc263d7cc2e369515f342957b76 |
|
I don't know what it means! |
|
If this backtrace doesn't enlighten the situation I'd suggest changing Ruby versions! |
|
Debian Stable is still at Ruby 2.3.3p222 while the testing branch currently runs 2.5.1. |
|
Here's a Ruby bug report with the same error message: https://bugs.ruby-lang.org/issues/11018
|
|
Hi @icarito, I have seen this error |
|
This is most likely a Gem with C-compiled extensions. Another way to try to catch it is with |
|
Notes from our work session: Todo: this file occurs a lot: https://github.com/ruby/ruby/blob/ruby_2_4/vm_insnhelper.c "invalid next size (fast)" related issues:
http://www.valgrind.org/ was suggested as an option to troubleshoot memory issues... |
|
Here's the relevant bit from Full trace (88MB) at https://gitlab.com/publiclab/plots2/-/jobs/88571931/artifacts/browse |
|
Still don't know what it means. Could it be related to #3086 ? :-/ |
|
I left a comment with the question at https://bugs.ruby-lang.org/issues/11018 - hopefully we'll get a response! |
|
Okay I've setup Jenkins to build/test every 10 minutes. We're seeing this error consistently. I've set it up to get us strace logs for each build and set string capture to 256 bytes. We can move things around and try to get it to stop failing / we can measure results now. |
|
The last strace had truncated strings. Here's the last 3000 lines. I think here's the more interesting bits: |
|
Having set to trace and test every 10 minutes, I have a lot of traces for this failure. |
|
There's no such gem as |
|
I wonder if they are built-in parts of ActiveSupport or another core Rails
component?
…On Thu, Aug 16, 2018 at 5:33 AM Emmanuel Hayford ***@***.***> wrote:
There's no such gem as dublincore on RubyGems and rss/maker looks strange
for a gem name @icarito <https://github.com/icarito>. Maybe they are not
gems after all?
—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub
<#2824 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AABfJ50F8UNW0VoZghVJeWVNdInXEh9bks5uRNm7gaJpZM4UjKqu>
.
|
|
Yes this component is part of Ruby proper. I've made the experiment of upgrading to Ruby 2.4.4 (we were at 2.4.1) - didn't fail all night testing every 10 minutes. This morning it began building |
ghost
assigned
|
Okay now It's building every 15 minutes hopefully will stay green, fixes in PR #3253 |
jywarren
pushed a commit
that referenced
this issue
Aug 27, 2018
* Added strace option and artifact * Restore Gitlab and add strace to Makefile * Use strace within container * Give more trace output * Switch invocation of passenger to fork 9 processes in testing * 256 string length. No op for triggering builds. * Try to speed up by not running bower / bundler * Build, then start. * Built before * Ignore logs in container * Try with newer docker ruby image * Set ruby version in Gemfile to 2.4.4 * Increase trace log to 512 * Don't export Redis port * Don't export web app port for testing * Set explicit Ruby version in Dockerfile * Avoid running bundle in Travis * Move strace to Gitlab to save disk space. * Add strace and artifacts to tests. * Remove redundant bundle and bower calls. * Add capability to be able to strace on Gitlab. * Run bower in Gitlab * Remove bash * Ignore tmp dir from docker container
SrinandanPai
pushed a commit
to SrinandanPai/plots2
that referenced
this issue
May 5, 2019
* Added strace option and artifact * Restore Gitlab and add strace to Makefile * Use strace within container * Give more trace output * Switch invocation of passenger to fork 9 processes in testing * 256 string length. No op for triggering builds. * Try to speed up by not running bower / bundler * Build, then start. * Built before * Ignore logs in container * Try with newer docker ruby image * Set ruby version in Gemfile to 2.4.4 * Increase trace log to 512 * Don't export Redis port * Don't export web app port for testing * Set explicit Ruby version in Dockerfile * Avoid running bundle in Travis * Move strace to Gitlab to save disk space. * Add strace and artifacts to tests. * Remove redundant bundle and bower calls. * Add capability to be able to strace on Gitlab. * Run bower in Gitlab * Remove bash * Ignore tmp dir from docker container
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Travis test sometimes gives this error
** Error in /usr/local/bin/ruby': free(): invalid next size (fast): 0x00000000098d3950 ***which gets cleared out when tests get rebuilded.The reported error is
There was an error with Danger bot’s Junit parsing: undefined methodvalue’ for nil:NilClass` in PRs.The text was updated successfully, but these errors were encountered: