mirror of
https://github.com/hoppscotch/hoppscotch.git
synced 2026-04-25 16:55:59 +03:00
[GH-ISSUE #3321] [bug]: Segmentation fault / Crashing Container when starting hoppscotch ce backend docker image on k8s #1104
Labels
No labels
CodeDay
a11y
browser limited
bug
bug fix
cli
core
critical
design
desktop
discussion
docker
documentation
duplicate
enterprise
feature
feature
fosshack
future
good first issue
hacktoberfest
help wanted
i18n
invalid
major
minor
need information
need testing
not applicable to hoppscotch
not reproducible
pull-request
question
refactor
resolved
sandbox
self-host
spam
stale
testmu
wip
wont fix
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
starred/hoppscotch#1104
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @RaHehl on GitHub (Sep 6, 2023).
Original GitHub issue: https://github.com/hoppscotch/hoppscotch/issues/3321
Originally assigned to: @balub on GitHub.
Is there an existing issue for this?
Current behavior
We are trying to move from our self-made images based on Webysther's PR to the new Docker hoppscotch/hoppscotch-backend:2023.8.0 image.
But our API container crashes with a segfault:
Steps to reproduce
use hoppscotch/hoppscotch-backend:2023.8.0
Environment
Production
Version
Self-hosted
@henrysachs commented on GitHub (Sep 6, 2023):
we were previously using the custom docker container mentioned a lot here in other issues. When taking a look into the old container logs we se that after pub sub normally:
graphql seems to boot. So I think there are some issues with the built container and the containing graphql library. Maybe a postinstall script or something like that.
@AndrewBastin commented on GitHub (Sep 6, 2023):
@henrysachs does this happen with
hoppscotch/hoppscotch:2023.8.0?@henrysachs commented on GitHub (Sep 6, 2023):
Yes
@AndrewBastin commented on GitHub (Sep 6, 2023):
We are unable to reproduce this on our end directly, so can we get some info on your environment and your k8s setup so we can try to repro it on our end ?
@henrysachs commented on GitHub (Sep 6, 2023):
its eks 1.23 with x86 nodes. We are deploying the independent containers. we also set a lot of environment variables but i think we cant provide all of them. Not sure what Info we can provide additionally. We are using helm to deploy.
@RaHehl commented on GitHub (Sep 7, 2023):
@AndrewBastin We got more information by loading the node coredump:
What reminded us that we had adapted the mailer module (packages/hoppscotch-backend/src/mailer/mailer.module.ts) in our custom build, to ignore ssl problems.
@henrysachs commented on GitHub (Sep 7, 2023):
so the fix seems to be upgrading the base docker image from alpine 3.16 to alpine 3.18
@henrysachs commented on GitHub (Sep 7, 2023):
these were our changes to make it work. Would you be open that we contribute this? :)
@AndrewBastin commented on GitHub (Sep 7, 2023):
@henrysachs thanks for the diff, we will look into this and verify the fixes, you can open up a PR for this. (cc: @balub, please verify segfaults that happen w.r.t Prisma).
The whole reason for the downgrade to Alpine 3.16 was because there is a weird sorta transition happening between the OpenSSL versions and how Node/native handle them. The updated OpenSSL version (v3) had lead us to have issues with Prisma crashing with a similar set of problems when the backend processes certain queries, which is why it is pinned down. Alpine 3.16 gives OpenSSL 1.1 on its repositories.
These issues (with OpenSSL) are seemed to be fixed in Node 20+ but since the LTS is still 18, we are waiting for the next LTS to hopefully clear these all up.
@AprilNEA commented on GitHub (Sep 14, 2023):
Same problem happened, hopefully it will be resolved soon.
@jordibeen commented on GitHub (Oct 2, 2023):
Having the same issue while trying to run it on Kubernetes, hoping the above mentioned fix will be rolled out soon!
@RaHehl commented on GitHub (Nov 21, 2023):
@AndrewBastin looks like 20 is now LTS? https://nodejs.org/en