sriov provider, run etcd in memory

Currently we encounter bad performance of etcd on sriov provider cluster on DinD setup. We get 'etcdserver: timeout errors' that causes jobs to fail often. In such cases it is recommanded [1] to use in-memory etcd Running etcd in memory should improve performance and will make sriov provider more stabilized. [1] kubernetes-sigs/kind#1922 Signed-off-by: Or Mergi <[email protected]>
ormergi · Nov 18, 2020 · e2bb958 · e2bb958
1 parent 38492d7
commit e2bb958
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 0 deletions.
diff --git a/cluster-up/cluster/kind-k8s-sriov-1.17.0/provider.sh b/cluster-up/cluster/kind-k8s-sriov-1.17.0/provider.sh
@@ -19,6 +19,9 @@ function up() {
 
     cp $KIND_MANIFESTS_DIR/kind.yaml ${KUBEVIRTCI_CONFIG_PATH}/$KUBEVIRT_PROVIDER/kind.yaml
 
+    export ETCD_DATA_DIR="/mnt/etcd"
+    ensure_memory_dir_for_etcd_data
+
     kind_up
 
     ${KUBEVIRTCI_PATH}/cluster/$KUBEVIRT_PROVIDER/config_sriov.sh

diff --git a/cluster-up/cluster/kind/common.sh b/cluster-up/cluster/kind/common.sh
@@ -270,7 +270,20 @@ EOF
     done
 }
 
+function _add_control_plane_extra_mounts() {
+    if [[ "$KUBEVIRT_PROVIDER" =~ sriov.* ]]; then
+        if [ "$ETCD_DATA_DIR" != none ]; then
+            cat <<EOF >> ${KUBEVIRTCI_CONFIG_PATH}/$KUBEVIRT_PROVIDER/kind.yaml
+  extraMounts:
+  - containerPath: /var/lib/etcd
+    hostPath: $ETCD_DATA_DIR
+EOF
+        fi
+    fi
+}
+
 function _prepare_kind_config() {
+    _add_control_plane_extra_mounts
     _add_workers
 
     echo "Final KIND config:"