Skip to content
Problems.

llama.cpp upgrades break working setups

Outages and change3 posts from 3 people3d active+33 posts in the last 7 days, 0 the 7 days before (new)

Posts per day

Posts per day3 posts, Sep 29 to Oct 5

The posts behind it

3, newest first
PostDate
Feature Request: gemma4 31B chat template quadratic scan cost### Prerequisites - [x] I am running the latest code. Mention the version if possible as well. - [x] I carefully followed the [README.md](https://github.com/ggml-org/llama.cpp/blob/master/README.md). - [x] I searched using keywords relevant to my issue to make sure that I am...ggml-org/llama.cpphasanyazarrOct 32 days ago
Feature Request: gracefully ignore missing/invalid arguments instead of failing the model loading in serve mode### Prerequisites - [x] I am running the latest code. Mention the version if possible as well. - [x] I carefully followed the [README.md](https://github.com/ggml-org/llama.cpp/blob/master/README.md). - [x] I searched using keywords relevant to my issue to make sure that I am...ggml-org/llama.cppCoderx7Oct 23 days ago
server: streaming responses advertise Keep-Alive but the connection is closed after the stream, causing ReadError/RemoteProtocolError on pooled HTTP clients...gaps (570 requests in one day, 7/283 vs 0/286). Under `strace` of llama-server, every failing request was on a reused connection, and the server called `shutdown(fd, SHUT_RDWR)` on that fd within about 1 ms of the client writing the new request. Every successful streamed request...ggml-org/llama.cppviiiviii538Sep 296 days ago

Companies and products named

Company or productPosts naming it
llama.cpp3
GitHub2
Hugging Face1
NVIDIA1
About this problem

Evidence

3 posts from 3 people in 1 place, about 3 a week over 5 days. Mostly on ggml-org/llama.cpp. Tools named alongside: GitHub, Hugging Face, NVIDIA.

Frustration Frustration 0 of 3· Seen on GitHub issues

How it was grouped

Posts that state a pain and match the "breaking-changes" rule about llama.cpp. First post Sep 29, 2026, latest Oct 3, 2026. Corroborated: 3 or more posts from 2 or more people or places. Method

History

  • 2026-10-05 Added: llama.cpp upgrades break working setups (3 posts)

Rising problems by email

Mondays: the problems in data, tech and AI that grew fastest that week.

Double opt-in. Unsubscribe any time.