straylight: raise OpenCode and llama-server timeouts to 2h

Local prefill and reasoning/coding turns often sit idle past OpenCode's
5–10 minute defaults. Align llama-server --timeout with that window.
This commit is contained in:
2026-09-17 05:49:52 -07:00
parent 2498ec066e
commit c5fed8c604
2 changed files with 8 additions and 1 deletions
+6
View File
@@ -26,6 +26,12 @@ let
options = { options = {
baseURL = llamaBase; baseURL = llamaBase;
apiKey = "none"; apiKey = "none";
# Local prefill + Ornith <think> / Laguna coding turns routinely sit
# quiet for >10 minutes. OpenCode defaults are 300000ms (request,
# headers, and inter-chunk). Milliseconds; false disables.
timeout = 7200000;
headerTimeout = 7200000;
chunkTimeout = 7200000;
}; };
models = { models = {
"laguna-s-2.1" = text "Laguna S 2.1"; "laguna-s-2.1" = text "Laguna S 2.1";
+2 -1
View File
@@ -187,7 +187,8 @@ in
--host 0.0.0.0 \ --host 0.0.0.0 \
--port 11434 \ --port 11434 \
--models-preset ${llamaModelsIni} \ --models-preset ${llamaModelsIni} \
--models-max 2 --models-max 2 \
--timeout 7200
''; '';
User = "llama-server"; User = "llama-server";
Group = "llama-server"; Group = "llama-server";