Out of memorry errors in CICD pipelines freqent for large third party dependencies #1662
Replies: 7 comments 2 replies
-
Could you include the log, so that we can determine whether you are using containers or compilerfolders? |
Beta Was this translation helpful? Give feedback.
-
alas I'm re-running the failed job(s) so I don't have a failed log to attach. if it fails again I'll attach the log., We are definately using containers however and NOT compilerfolders wer'e aware that LS Central for BC has a huge memory footprint (both onprem and saas) which has presented its own chalenges in the past when too many onprem instances on a server have LS installed and onprem BC shell experiences frequent out of memory errors if there are too many instances with LS installed., |
Beta Was this translation helpful? Give feedback.
-
In github, you can select the prior re-run in the top right corner and see the log from the failing builds. |
Beta Was this translation helpful? Give feedback.
-
duly noted! |
Beta Was this translation helpful? Give feedback.
-
while I wait for the CICD to finish with the memorylimit set to 16G, I found an older log that has the out of memory exception. see attached. in the log zip it's the "CA" project that failed @ build step with out of memory |
Beta Was this translation helpful? Give feedback.
-
after setting memoryLimit to "16G" didn't help. the pipline still randomly throws out-of-memory errors during the build step, requiring re-running the failed workflow step 3-4x before succeeding. below is a copy/paste of the log when it fails installing a large third party app
and further down the log I see this:
@freddydk I asked copilot about the memorylimit=16G setting and it suggested to reduce the value to 14G due windows-based github-runner VM overhead (not that II think it'll help in this case as evidenced by the it would appear theres only two workarounds to this error:
thoughts?? |
Beta Was this translation helpful? Give feedback.
-
@freddydk thanks for the confirmation. I'm pivoting to self-hosted runners this week to prove it out (not all our runners need to be self-hosted but the ones that depend on LS Retail do) |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
AL-Go version
6.3
Describe the issue
we're currently building an appsource app that exends appsource app LS Central (Publisher: LS Retail).
50% of the time CICD fails at the build step due to out of memory errors (LS for BC takes a large resource footprint). Currently we re-run the failed job(s) and cross our fingers.
Short of exploring self-hosted runners, is there anything that can be done to improve CICD reliability?
Expected behavior
CICD succeeds without any out of memory errors
Steps to reproduce
create an app that depends on LS Central
Additional context (logs, screenshots, etc.)
No response
Beta Was this translation helpful? Give feedback.
All reactions