| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
remove hardcoded eks cluster name Signed-off-by: coloryourlife <95chris0319@gmail.com> | 1 年前 | |
[Tutorial] Deployment on Azure AKS (#247) * [CI]: Add a ignore file for codespell This commit adds a .codespell-ignore that can be used to ignore the technical words and abbreviations and anything else that is not part of the standard dictionary, but relevant in the context of this repo. Signed-off-by: Suraj Deshmukh <surajd.service@gmail.com> * [Tutorial] Deployment on Azure AKS This commit adds docs and scripts to deploy the production stack on Azure AKS. Signed-off-by: Suraj Deshmukh <surajd.service@gmail.com> --------- Signed-off-by: Suraj Deshmukh <surajd.service@gmail.com> | 1 年前 | |
[bugfix] deprecate disable log request (#885) * bugfix: deprecate disable log request Signed-off-by: Rui Zhang <zrfishnoodles@gmail.com> * Update helm/README.md Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Signed-off-by: Rui Zhang <51696593+ruizhang0101@users.noreply.github.com> --------- Signed-off-by: Rui Zhang <zrfishnoodles@gmail.com> Signed-off-by: Rui Zhang <51696593+ruizhang0101@users.noreply.github.com> Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> | 5 个月前 | |
Fix OCI OKE deployment script (entry_point.sh) — end-to-end tested (#811) * fix(oci): refactor entry_point.sh and align README with end-to-end tested deployment Key changes to entry_point.sh: - Refactor expand_gpu_disk to single kubectl exec with retry logic - Replace kubectl wait -l app=gpu with polling loop (handles no-match) - Add || true guards for kubectl exec exit code 137 (SIGKILL after success) - Fix set -euo pipefail compatibility when no GPU nodes exist yet - Use kubectl config view --minify -o jsonpath for kubeconfig cluster name - Remove --wait from helm deploy (router CrashLoops until patched) - Use nsenter instead of chroot for kubelet restart (systemd bus access) - Add CPU_BOOT_VOLUME_GB=100 to prevent DiskPressure from router image - Revert image queries to !contains('aarch64') — OKE images lack x86_64 - Cleanup deletes all clusters with matching name, not just the first - Fix stdout pollution breaking filesystem size verification Key changes to README.md: - Use kubectl config view --minify for kubeconfig (matches entry_point.sh) - Replace separate restart-kubelet pod with nsenter via expand-gpu-disk pod - Add CPU_BOOT_VOLUME_GB to environment variables table - Document 4 additional gotchas (exit code 137, CPU disk pressure, OCI image naming, kubeconfig context vs cluster name) Signed-off-by: Federico Kamelhar <federico.kamelhar@oracle.com> * docs: add OCI security callouts Signed-off-by: Federico Kamelhar <federico.kamelhar@oracle.com> --------- Signed-off-by: Federico Kamelhar <federico.kamelhar@oracle.com> Co-authored-by: Rui Zhang <51696593+ruizhang0101@users.noreply.github.com> | 6 个月前 |