Ensure ABT_THREAD_STACKSIZE>=18432 for dcpm storage class

dcpm storage class with NDCTL enabled requires at least 18432 stack size for ULT. PR-repos: pmdk@PR-38:11 Priority: 2 Cancel-prev-build: false Skip-nlt: false Skip-unit-tests: false Skip-func-test-vm: false Skip-test-rpms: false Allow-unstable-test: true Skip-func-hw-test-large: false Skip-func-hw-test-medium: false Skip-func-hw-test-medium-verbs-provider: true Required-githooks: true Signed-off-by: Tomasz Gromadzki <[email protected]>
daos-stack · Aug 28, 2024 · f74f025 · f74f025
1 parent ee6d615
commit f74f025
Show file tree

Hide file tree

Showing 3 changed files with 37 additions and 0 deletions.
diff --git a/src/control/cmd/dmg/auto_test.go b/src/control/cmd/dmg/auto_test.go
@@ -566,6 +566,8 @@ engines:
   provider: ofi+verbs
   fabric_iface: ib0
   fabric_iface_port: 31416
+  env_vars:
+  - ABT_THREAD_STACKSIZE=18432
   pinned_numa_node: 0
 - targets: 6
   nr_xs_helpers: 0
@@ -583,6 +585,8 @@ engines:
   provider: ofi+verbs
   fabric_iface: ib1
   fabric_iface_port: 32416
+  env_vars:
+  - ABT_THREAD_STACKSIZE=18432
   pinned_numa_node: 1
 disable_vfio: false
 disable_vmd: false

diff --git a/src/control/server/engine/config.go b/src/control/server/engine/config.go
@@ -9,6 +9,7 @@ package engine
 import (
 	"fmt"
 	"os"
+	"strconv"
 	"strings"
 
 	"github.com/pkg/errors"
@@ -28,6 +29,8 @@ const (
 	envLogMasks      = "D_LOG_MASK"
 	envLogDbgStreams = "DD_MASK"
 	envLogSubsystems = "DD_SUBSYS"
+
+	MIN_ABT_THREAD_STACKSIZE_FOR_DCPM = 18432
 )
 
 // FabricConfig encapsulates networking fabric configuration.
@@ -343,6 +346,23 @@ func (c *Config) Validate() error {
 		return errors.Wrap(err, "validate engine log subsystems")
 	}
 
+	// ensure 18KB ABT stack size for an engine with DCPM storage class
+	if c.Storage.Tiers[0].Class == storage.ClassDcpm {
+		stacksize_str, err := c.GetEnvVar("ABT_THREAD_STACKSIZE")
+		if err != nil {
+			fmt.Printf("env_var ABT_THREAD_STACKSIZE set to %d for `dcpm` storage class\n",
+				MIN_ABT_THREAD_STACKSIZE_FOR_DCPM)
+			c.EnvVars = append(c.EnvVars, fmt.Sprintf("ABT_THREAD_STACKSIZE=%d",
+				MIN_ABT_THREAD_STACKSIZE_FOR_DCPM))
+		} else {
+			stacksize_value, err := strconv.Atoi(stacksize_str)
+			if err != nil || stacksize_value < MIN_ABT_THREAD_STACKSIZE_FOR_DCPM {
+				return errors.New(fmt.Sprintf("env_var ABT_THREAD_STACKSIZE should be >= %d for `dcpm` storage class",
+					MIN_ABT_THREAD_STACKSIZE_FOR_DCPM))
+			}
+		}
+	}
+
 	return nil
 }
 
@@ -480,6 +500,10 @@ func (c *Config) WithSystemName(name string) *Config {
 func (c *Config) WithStorage(cfgs ...*storage.TierConfig) *Config {
 	c.Storage.Tiers = storage.TierConfigs{}
 	c.AppendStorage(cfgs...)
+	if len(c.Storage.Tiers) > 0 && c.Storage.Tiers[0].IsSCM() &&
+		c.Storage.Tiers[0].Class == storage.ClassDcpm {
+		return c.WithStackSizeForDCPM()
+	}
 	return c
 }
 
@@ -690,3 +714,9 @@ func (c *Config) WithStorageIndex(i uint32) *Config {
 	c.Storage.EngineIdx = uint(i)
 	return c
 }
+
+func (c *Config) WithStackSizeForDCPM() *Config {
+	c.EnvVars = append(c.EnvVars, fmt.Sprintf("ABT_THREAD_STACKSIZE=%d",
+		MIN_ABT_THREAD_STACKSIZE_FOR_DCPM))
+	return c
+}
diff --git a/src/tests/ftest/util/server_utils_params.py b/src/tests/ftest/util/server_utils_params.py
@@ -537,6 +537,9 @@ def get_params(self, test):
 
         if not self.using_dcpm:
             required_env_vars["PMEMOBJ_CONF"] = "sds.at_create=0"
+        else:
+            # NDCTL enabling requires bigger ABT default stack size
+            required_env_vars["ABT_THREAD_STACKSIZE"] = "18432"
 
         # Enable fault injection if configured
         if test.fault_injection.fault_file is not None: