Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-19 14:13:35 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #4365] [BUG]: node process is killed while uploading and embedding documents #2784

New Issue
Closed
opened 2026-02-22 18:31:14 -05:00 by yindo · 2 comments
yindo commented 2026-02-22 18:31:14 -05:00
Owner
Copy Link
Copy Source

Originally created by @jstawski on GitHub (Sep 5, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4365

How are you running AnythingLLM?

Docker (remote machine)

What happened?

I am running the PG container in a hosting environment. The container has 1 CPU and 3 GB of memory. When uploading a document and embedding it directly into a workspace (utilizing the API), the node process is sometimes killed. I haven't pinpointed if this happens with only one file or if it happens after a series of files (does the embedding happen in a background process?). Here's an excerpt of the logs:

[36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 76 of 406
[36m[backend][0m [32minfo[39m: [36m[CollectorApi][0m Document Office filename.json uploaded, processed, and moved to folder Workspace1 successfully.
[36m[backend][0m [32minfo[39m: [32m[Event Logged][0m - api_document_uploaded
[36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 393 of 406
[36m[backend][0m [32minfo[39m: [35m[PGVectorDb][0m Adding new vectorized document into namespace workspace1
[36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Initialized Xenova/all-MiniLM-L6-v2
[36m[backend][0m [32minfo[39m: [35m[RecursiveSplitter][0m Will split with {"chunkSize":1000,"chunkOverlap":20,"chunkHeader":"<document_metadata>\nsourceDocument: Office HoursYo..."}
[36m[backend][0m [32minfo[39m: [35m[PGVectorDb][0m Snippets created from document: 10133
[36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 77 of 406
[36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 394 of 406
...
[36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 79 of 406
/usr/local/bin/docker-entrypoint.sh: line 24: 162 Killed node /app/server/index.js

Looking at the metrics of the container I can see big CPU and Memory usage during this time, which eventually goes down to 0 (since the process was killed). In the memory side of things the memory is moving up and down between ~1.7 GB and ~2.4 GB, eventually it spikes to 3GB and dies.

Everything points to an Out of Memory exception, but I can't find any logs besides the 162 Killed node message.

Am I running with too little memory? If I double it, will this continue to happen, just with a bigger load? How can I avoid it?

Are there known steps to reproduce?

Setup a container with constrained memory and upload many docs of different sizes and embed them. I was able to reproduce via the API and the UI.

Originally created by @jstawski on GitHub (Sep 5, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/4365 ### How are you running AnythingLLM? Docker (remote machine) ### What happened? I am running the PG container in a hosting environment. The container has 1 CPU and 3 GB of memory. When uploading a document and embedding it directly into a workspace (utilizing the API), the node process is sometimes killed. I haven't pinpointed if this happens with only one file or if it happens after a series of files (does the embedding happen in a background process?). Here's an excerpt of the logs: > [36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 76 of 406 [36m[backend][0m [32minfo[39m: [36m[CollectorApi][0m Document Office filename.json uploaded, processed, and moved to folder Workspace1 successfully. [36m[backend][0m [32minfo[39m: [32m[Event Logged][0m - api_document_uploaded [36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 393 of 406 [36m[backend][0m [32minfo[39m: [35m[PGVectorDb][0m Adding new vectorized document into namespace workspace1 [36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Initialized Xenova/all-MiniLM-L6-v2 [36m[backend][0m [32minfo[39m: [35m[RecursiveSplitter][0m Will split with {"chunkSize":1000,"chunkOverlap":20,"chunkHeader":"<document_metadata>\nsourceDocument: Office HoursYo..."} [36m[backend][0m [32minfo[39m: [35m[PGVectorDb][0m Snippets created from document: 10133 [36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 77 of 406 [36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 394 of 406 ... [36m[backend][0m [32minfo[39m: [36m[NativeEmbedder][0m Embedded Chunk Group 79 of 406 /usr/local/bin/docker-entrypoint.sh: line 24: 162 Killed node /app/server/index.js Looking at the metrics of the container I can see big CPU and Memory usage during this time, which eventually goes down to 0 (since the process was killed). In the memory side of things the memory is moving up and down between ~1.7 GB and ~2.4 GB, eventually it spikes to 3GB and dies. Everything points to an Out of Memory exception, but I can't find any logs besides the `162 Killed node` message. Am I running with too little memory? If I double it, will this continue to happen, just with a bigger load? How can I avoid it? ### Are there known steps to reproduce? Setup a container with constrained memory and upload many docs of different sizes and embed them. I was able to reproduce via the API and the UI.
yindo added the possible bug label 2026-02-22 18:31:14 -05:00
yindo closed this issue 2026-02-22 18:31:14 -05:00
yindo commented 2026-02-22 18:31:14 -05:00
Author
Owner
Copy Link
Copy Source

@shatfield4 commented on GitHub (Sep 5, 2025):

Yes, this does seem like it is an out of memory (OOM) exception. I can see that the document you are uploading is rather large since it is embedding 406 chunks. Try increasing the CPU and RAM allocated to the docker container and this may solve the issue.

You can confirm that this is the issue by trying a smaller document on the same setup and if it completes properly then it is indeed an OOM exception.

@shatfield4 commented on GitHub (Sep 5, 2025): Yes, this does seem like it is an out of memory (OOM) exception. I can see that the document you are uploading is rather large since it is embedding 406 chunks. Try increasing the CPU and RAM allocated to the docker container and this may solve the issue. You can confirm that this is the issue by trying a smaller document on the same setup and if it completes properly then it is indeed an OOM exception.
yindo commented 2026-02-22 18:31:14 -05:00
Author
Owner
Copy Link
Copy Source

@jstawski commented on GitHub (Sep 8, 2025):

But why is the process killed? Isn't there a way to make sure it doesn't die?

@jstawski commented on GitHub (Sep 8, 2025): But why is the process killed? Isn't there a way to make sure it doesn't die?
yindo changed title from [BUG]: node process is killed while uploading and embedding documents to [GH-ISSUE #4365] [BUG]: node process is killed while uploading and embedding documents 2026-06-05 14:48:32 -04:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Desktop
Docker
Integration Request
Integration Request
OS: Linux
OS: Mobile
OS: Windows
UI/UX
blocked
bug
bug
core-team-only
documentation
duplicate
embed-widget
enhancement
feature request
github_actions
good first issue
investigating
needs info / can't replicate
possible bug
pull-request
Mirrored from GitHub Pull Request
 question
stage: specifications
wontfix
No Label possible bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#2784
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?