hpcugent · itkovian · Mar 13, 2025 · Mar 13, 2025 · Mar 13, 2025 · Mar 13, 2025
diff --git a/META b/META
@@ -7,8 +7,8 @@
   Name:		slurm
   Major:	24
   Minor:	11
-  Micro:	3
-  Version:	24.11.3
+  Micro:	5
+  Version:	24.11.5
   Release:	1
 
 ##

diff --git a/NEWS b/NEWS
diff --git a/build.sh b/build.sh
@@ -18,7 +18,7 @@ SCRIPT=$(readlink -f "${BASH_SOURCE[0]}")
 ORIGIN=$(dirname "$SCRIPT")
 
 # which version to download from github
-SLURM_VERSION=${VERSION:-24.05.3}
+SLURM_VERSION=${VERSION:-24.11.4}
 UPSTREAM_REL=${UPSTREAM_REL:-1}
 
 # which release should be used for our RPMs
@@ -28,17 +28,24 @@ OUR_RELEASE=${RELEASE:-1}
 # allow _empty_ version, which is used in pipeline
 
 if grep "release 8.8" /etc/redhat-release; then
-   NVIDIA_DRIVER=${NVIDIA_DRIVER-555.42.06}
-   NVDRV_NVML_PKG="nvidia-driver-NVML${NVIDIA_DRIVER:+-$NVIDIA_DRIVER}"
-   CUDA_VERSION=${CUDA_VERSION:-12.6}
-   CUDA_NVML_PKG="cuda-nvml-devel-${CUDA_VERSION//./-}"
+    NVIDIA_MAJOR_VERSION=570
+    NVIDIA_MINOR_VERSION=133.20
+    NVIDIA_DRIVER=${NVIDIA_DRIVER-${NVIDIA_MAJOR_VERSION}.${NVIDIA_MINOR_VERSION}}
+    NVDRV_NVML_PKG="libnvidia-ml${NVIDIA_DRIVER:+-$NVIDIA_DRIVER}"
+    CUDA_VERSION=${CUDA_VERSION:-12.8}
+    CUDA_NVML_PKG="cuda-nvml-devel-${CUDA_VERSION//./-}"
 elif grep "release 9.4" /etc/redhat-release; then
-   NVIDIA_DRIVER=${NVIDIA_DRIVER-555.42.06}
-   NVDRV_NVML_PKG="nvidia-driver-NVML${NVIDIA_DRIVER:+-$NVIDIA_DRIVER}"
-   CUDA_VERSION=${CUDA_VERSION:-12.6}
-   CUDA_NVML_PKG="cuda-nvml-devel-${CUDA_VERSION//./-}"
+    NVIDIA_MAJOR_VERSION=570
+    NVIDIA_MINOR_VERSION=133.20
+    NVIDIA_DRIVER=${NVIDIA_DRIVER-${NVIDIA_MAJOR_VERSION}.${NVIDIA_MINOR_VERSION}}
+    NVDRV_NVML_PKG="libnvidia-ml${NVIDIA_DRIVER:+-$NVIDIA_DRIVER}"
+    CUDA_VERSION=${CUDA_VERSION:-12.8}
+    CUDA_NVML_PKG="cuda-nvml-devel-${CUDA_VERSION//./-}"
 fi
 
+
+
+
 # Prepare directory structure
 rm -Rf $ORIGIN/rpmbuild/ $ORIGIN/dist/
 mkdir -p $ORIGIN/rpmbuild/{BUILD,RPMS,SRPMS,SOURCES} $ORIGIN/dist
@@ -97,7 +104,10 @@ sudo dnf -y install munge-devel libjwt-devel pam-devel
 sudo dnf -y install http-parser-devel json-c-devel libyaml-devel
 # - features: Nvidia NVML
 sudo dnf -y autoremove cuda-nvml-* nvidia-driver-NVML-* nvidia-driver* libnvidia-ml*
-sudo dnf -y install "$CUDA_NVML_PKG" "$NVDRV_NVML_PKG" "nvidia-driver-devel"
+
+sudo dnf -y module switch-to nvidia-driver:${NVIDIA_MAJOR_VERSION}-dkms
+
+sudo dnf -y install "$CUDA_NVML_PKG" "$NVDRV_NVML_PKG" # "nvidia-driver-devel"
 # - plugins: MPI
 sudo dnf -y install pmix "pmix-devel ${PMIX_VERSION}"  "ucx-devel-${UCX_VERSION}"
 # - plugins: cgroup/v2
@@ -128,7 +138,7 @@ rpmbuild -ba "${RPM_DEFINES[@]}" "${SLURM_BUILDOPTS[@]}" --without nvml \
 echo "Doing rpm rebuild (without nvml)"
 for rpm in $ORIGIN/rpmbuild/RPMS/x86_64/slurm-*$SUFFIX*.rpm ; do
     rpmrebuild --release=${OUR_RELEASE}.${GITTAG}$(rpm -E '%dist').nogpu.ug -d $ORIGIN/dist -p $rpm
-done
+done 2>&1 | tee rpmrebuild-without-nvml.out
 
 
 echo "Running rpmbuild (with nvml)"
@@ -139,7 +149,7 @@ rpmbuild -ba "${RPM_DEFINES[@]}" "${SLURM_BUILDOPTS[@]}" --with nvml \
 echo "Doing rpm rebuild (with nvml)"
 for rpm in $ORIGIN/rpmbuild/RPMS/x86_64/slurm-*$SUFFIX*.rpm ; do
     rpmrebuild --release=${OUR_RELEASE}.${GITTAG}$(rpm -E '%dist').ug -d $ORIGIN/dist -p $rpm
-done
+done 2>&1 | tee rpmrebuild-with-nvml.out
 
 # strip out torque binaries/wrapper from slurm-torque
 rpmrebuild -d $ORIGIN/dist --change-spec-files="sed '/\(pbsnodes\|mpiexec\|bin\/q.\+\)/d'" -p $ORIGIN/dist/x86_64/slurm-torque-*-${OUR_RELEASE}.${GITTAG}$(rpm -E '%dist').nogpu.ug*.rpm

diff --git a/cve-2025-43904_24.05_24.11.patch b/cve-2025-43904_24.05_24.11.patch
@@ -0,0 +1,72 @@
+diff --git a/src/plugins/accounting_storage/mysql/as_mysql_user.c b/src/plugins/accounting_storage/mysql/as_mysql_user.c
+index 4d337be15b..dc73344783 100644
+--- a/src/plugins/accounting_storage/mysql/as_mysql_user.c
++++ b/src/plugins/accounting_storage/mysql/as_mysql_user.c
+@@ -542,6 +542,7 @@ extern int as_mysql_add_users(mysql_conn_t *mysql_conn, uint32_t uid,
+ 	if (check_connection(mysql_conn) != SLURM_SUCCESS)
+ 		return ESLURM_DB_CONNECTION;
+
++	bool is_admin = false;
+ 	if (!is_user_min_admin_level(mysql_conn, uid, SLURMDB_ADMIN_OPERATOR)) {
+ 		slurmdb_user_rec_t user;
+
+@@ -563,6 +564,8 @@ extern int as_mysql_add_users(mysql_conn_t *mysql_conn, uint32_t uid,
+ 		 * these accounts if they are coordinators of the
+ 		 * parent they are trying to add to
+ 		 */
++	} else {
++		is_admin = true;
+ 	}
+
+ 	if (!user_list || !list_count(user_list)) {
+@@ -588,6 +591,11 @@ extern int as_mysql_add_users(mysql_conn_t *mysql_conn, uint32_t uid,
+ 			   (long)now, (long)now, object->name);
+
+ 		if (object->admin_level != SLURMDB_ADMIN_NOTSET) {
++			if (!is_admin) {
++				error("Only admins/operators can add make a user and operator/admin");
++				rc = ESLURM_ACCESS_DENIED;
++				break;
++			}
+ 			xstrcat(cols, ", admin_level");
+ 			xstrfmtcat(vals, ", %u", object->admin_level);
+ 			xstrfmtcat(extra, ", admin_level=%u",
+@@ -669,7 +677,7 @@ extern int as_mysql_add_users(mysql_conn_t *mysql_conn, uint32_t uid,
+ 	list_iterator_destroy(itr);
+ 	xfree(user_name);
+
+-	if (rc != SLURM_ERROR) {
++	if (rc == SLURM_SUCCESS) {
+ 		if (txn_query) {
+ 			xstrcat(txn_query, ";");
+ 			rc = mysql_db_query(mysql_conn,
+@@ -714,7 +722,16 @@ extern char *as_mysql_add_users_cond(mysql_conn_t *mysql_conn, uint32_t uid,
+ 	}
+
+ 	if (!is_user_min_admin_level(mysql_conn, uid, SLURMDB_ADMIN_OPERATOR)) {
+-		slurmdb_user_rec_t user;
++		slurmdb_user_rec_t user_coord = {
++			.uid = uid,
++		};
++
++		if (user->admin_level != SLURMDB_ADMIN_NOTSET) {
++			ret_str = xstrdup("Only admins/operators can add make a user and admin/operator");
++			error("%s", ret_str);
++			errno = ESLURM_ACCESS_DENIED;
++			return ret_str;
++		}
+
+ 		if (slurmdbd_conf->flags & DBD_CONF_FLAG_DISABLE_COORD_DBD) {
+ 			ret_str = xstrdup("Coordinator privilege revoked with DisableCoordDBD, only admins/operators can add accounts.");
+@@ -723,10 +740,7 @@ extern char *as_mysql_add_users_cond(mysql_conn_t *mysql_conn, uint32_t uid,
+ 			return ret_str;
+ 		}
+
+-		memset(&user, 0, sizeof(slurmdb_user_rec_t));
+-		user.uid = uid;
+-
+-		if (!is_user_any_coord(mysql_conn, &user)) {
++		if (!is_user_any_coord(mysql_conn, &user_coord)) {
+ 			ret_str = xstrdup("Only admins/operators/coordinators can add accounts");
+ 			error("%s", ret_str);
+ 			errno = ESLURM_ACCESS_DENIED;
diff --git a/debian/changelog b/debian/changelog
@@ -1,4 +1,4 @@
-slurm-smd (24.11.3-1) UNRELEASED; urgency=medium
+slurm-smd (24.11.5-1) UNRELEASED; urgency=medium
 
   * Initial release.
 

diff --git a/doc/html/accounting.shtml b/doc/html/accounting.shtml
@@ -460,7 +460,7 @@ SlurmDBD executes</li>
 
 <li><b>AccountingStoragePass</b>: The password used to gain access to the
 database to store the accounting data. Only used for database type storage
-plugins, ignored otherwise. In the case of Slurm DBD (Database Daemon) with
+plugins, ignored otherwise. In the case of SlurmDBD (Database Daemon) with
 MUNGE authentication this can be configured to use a MUNGE daemon specifically
 configured to provide authentication between clusters while the default MUNGE
 daemon provides authentication within a cluster. In that case,

diff --git a/doc/html/cgroups.shtml b/doc/html/cgroups.shtml
@@ -217,10 +217,13 @@ details.</p>
 specific amount of memory on each compute node for exclusive use by the Slurm
 compute node daemon, slurmd.</p>
 
-<p>Slurmstepd is not constrained by this resource specialization, since it is
-considered part of the job and its consumption is completely dependent on the
-typology of the job. For example an MPI job can initialize many ranks with PMI
-and make slurmstepd consume more memory.</p>
+<p>If cgroup/v1 is used the reserved resources will also be used by the
+slurmstepd processes. If cgroup/v2 is used, slurmstepd is not constrained by
+this resource specialization. Instead the slurmstepd is constrained to the
+resources allocated to the job, since it is considered part of the job and its
+consumption is completely dependent on the topology of the job. For example an
+MPI job can initialize many ranks with PMI and make slurmstepd consume more
+memory.</p>
 
 <p>System-level resource specialization is enabled with special node
 configuration parameters. Read <a href="slurm.conf.html">slurm.conf</a> and core
@@ -421,6 +424,6 @@ controllers are under separate directories when using
 </li>
 </ul>
 
-<p style="text-align:center;">Last modified 11 October 2024</p>
+<p style="text-align:center;">Last modified 4 April 2025</p>
 
 <!--#include virtual="footer.txt"-->
diff --git a/doc/html/configless_slurm.shtml b/doc/html/configless_slurm.shtml
@@ -20,8 +20,8 @@ from, those client commands will have to use the DNS record to get the
 configuration information from the controller when they run.
 If you expect to have a lot of traffic from a login node, this
 can generate a lot of requests for the configuration files. In cases like
-this, you may want to consider running slurmd on the machine so it can
-manage the configuration files, but not allowing it to run jobs.</p>
+this, <a href="sackd.html">sackd</a> can be used to manage configuration files
+for the node reducing network requests.</p>
 
 <h2 id="INSTALLATION">Installation
 <a class="slurm_link" href="#INSTALLATION"></a>
@@ -148,6 +148,6 @@ is as follows:</p>
 <li>topology.conf</li>
 </ul>
 
-<p style="text-align:center;">Last modified 31 October 2024</p>
+<p style="text-align:center;">Last modified 18 April 2025</p>
 
 <!--#include virtual="footer.txt"-->
diff --git a/doc/html/configurator.easy.html.in b/doc/html/configurator.easy.html.in
@@ -437,7 +437,7 @@ Select one value for <B>AccountingStorageType</B>:<BR>
 <input type="radio" name="accounting_storage_type" value="" checked> <B>None</B>:
 No job accounting storage<BR>
 <input type="radio" name="accounting_storage_type" value="accounting_storage/slurmdbd"> <B>SlurmDBD</B>:
-Write job accounting to Slurm DBD (database daemon) which can securely
+Write job accounting to SlurmDBD (database daemon) which can securely
 save the data from many Slurm managed clusters into a common database<BR>
 <P>
 

diff --git a/doc/html/configurator.html.in b/doc/html/configurator.html.in
@@ -613,7 +613,7 @@ Select one value for <B>AccountingStorageType</B>:<BR>
 <input type="radio" name="accounting_storage_type" value="" checked> <B>None</B>:
 No job accounting storage<BR>
 <input type="radio" name="accounting_storage_type" value="accounting_storage/slurmdbd"> <B>SlurmDBD</B>:
-Write job accounting to Slurm DBD (database daemon) which can securely
+Write job accounting to SlurmDBD (database daemon) which can securely
 save the data from many Slurm managed clusters into a common database<BR>
 <input type="text" name="accounting_storage_loc" value=""> <B>AccountingStorageLoc</B>:
 Location specification or database name.

diff --git a/doc/html/containers.shtml b/doc/html/containers.shtml
@@ -857,20 +857,20 @@ Minimalist Dockerfile to generate a image with OpenMPI and PMIx to test basic MP
 <h4>Dockerfile</h4>
 <pre>
 FROM almalinux:latest
-RUN dnf -y update && dnf -y upgrade && dnf install -y yum-utils && dnf config-manager --set-enabled powertools
-RUN dnf -y install make automake gcc gcc-c++ kernel-devel bzip2 python3 wget libevent-devel hwloc-devel munge-devel
+RUN dnf -y update && dnf -y upgrade && dnf install -y epel-release && dnf -y update
+RUN dnf -y install make automake gcc gcc-c++ kernel-devel bzip2 python3 wget libevent-devel hwloc-devel
 
 WORKDIR /usr/local/src/
-RUN wget 'https://github.com/openpmix/openpmix/releases/download/v4.2.2/pmix-4.2.2.tar.bz2' -O - | tar -xvjf -
-WORKDIR /usr/local/src/pmix-4.2.2/
+RUN wget --quiet 'https://github.com/openpmix/openpmix/releases/download/v5.0.7/pmix-5.0.7.tar.bz2' -O - | tar --no-same-owner -xvjf -
+WORKDIR /usr/local/src/pmix-5.0.7/
 RUN ./configure && make -j && make install
 
 WORKDIR /usr/local/src/
-RUN wget --inet4-only 'https://download.open-mpi.org/release/open-mpi/v5.0/openmpi-5.0.0rc9.tar.gz' -O - | tar -xvzf -
-WORKDIR /usr/local/src/openmpi-5.0.0rc9
+RUN wget --quiet --inet4-only 'https://download.open-mpi.org/release/open-mpi/v5.0/openmpi-5.0.7.tar.bz2' -O - | tar --no-same-owner -xvjf -
+WORKDIR /usr/local/src/openmpi-5.0.7/
 RUN ./configure --disable-pty-support --enable-ipv6 --without-slurm --with-pmix --enable-debug && make -j && make install
 
-WORKDIR /usr/local/src/openmpi-5.0.0rc9/examples
+WORKDIR /usr/local/src/openmpi-5.0.7/examples
 RUN make && cp -v hello_c ring_c connectivity_c spc_example /usr/local/bin
 </pre>
 

diff --git a/doc/html/core_spec.shtml b/doc/html/core_spec.shtml
@@ -125,7 +125,10 @@ Specialized resources should be configured in slurm.conf on the
 node specification line using the <i>CoreSpecCount</i> or <i>CPUSpecList</i>
 options to identify the CPUs to reserve.
 The <i>MemSpecLimit</i> option can be used to reserve memory.
-These resources will be reserved using Linux cgroups.
+These resources will be reserved using Linux cgroups. The compute node daemon,
+slurmd, will be constrained to the reserved resources unless
+<i>TaskPluginParam</i> <i>SlurmdOffSpec</i> is specified. If cgroup/v1 is used,
+the slurmstepd processes will also be constrained to the reserved resources.
 Users wanting a different number of specialized cores should use the
 <i>--core-spec</i> option as described above.</p>
 
@@ -141,6 +144,6 @@ system use.</p>
 cores for their job, the configuration parameter <i>AllowSpecResourcesUsage</i>
 must be set to a value of <i>1</i>.</p>
 
-<p style="text-align:center;">Last modified 16 January 2024</p>
+<p style="text-align:center;">Last modified 4 April 2025</p>
 
 <!--#include virtual="footer.txt"-->
diff --git a/doc/html/faq.shtml b/doc/html/faq.shtml
@@ -134,6 +134,7 @@
   expected level?</a></li>
 <li><a href="#delete_partition">How do I safely remove partitions?</a></li>
 <li><a href="#routing_queue">How can a routing queue be configured?</a></li>
+<li><a href="#none_plugins">What happened to the "none" plugins?</a></li>
 </ul>
 <h3>Accounting Database</h3>
 <ul>
@@ -1372,6 +1373,35 @@ guide for more information.
 Also see the available job submit plugins distributed with Slurm for examples
 (look in the "src/plugins/job_submit" directory).</p>
 
+<p><a id="none_plugins"><b>What happened to the "none" plugins?</b></a><br>
+In Slurm 23.02 and earlier, several parameters had a plugin named "none"
+that would essentially disable the setting. In version 23.11, those plugins
+named "none" were removed. To disable a setting you just need to leave it
+unset. If you still have a plugin defined as "none", Slurm will still
+recognize it and treat it as though it was unset. Parameters that previously
+had a "none" plugin are:
+<ul>
+<li>AccountingStorageType</li>
+<li>AcctGatherEnergyType</li>
+<li>AcctGatherInterconnectType</li>
+<li>AcctGatherFilesystemType</li>
+<li>AcctGatherProfileType</li>
+<li>CliFilterPlugins</li>
+<li>CoreSpecPlugin</li>
+<li>ExtSensorsType</li>
+<li>JobAcctGatherType</li>
+<li>JobCompType</li>
+<li>JobContainerType</li>
+<li>MCSPlugin</li>
+<li>MpiDefault</li>
+<li>PowerParameters</li>
+<li>PreemptType</li>
+<li>PrioritySiteFactorPlugin</li>
+<li>SwitchType</li>
+<li>TaskPlugin</li>
+<li>TopologyPlugin</li>
+</ul></p>
+
 <h3>Accounting Database</h3>
 
 <p><a id="slurmdbd"><b>Why should I use the slurmdbd instead of the

diff --git a/doc/html/openapi_release_notes.shtml b/doc/html/openapi_release_notes.shtml
@@ -49,6 +49,10 @@ any target slurmrestd daemon. Development of clients should always be designed
 to use the highest version of the plugins available to avoid needing to port
 clients sooner than would be otherwise required.</p>
 
+<!--#include virtual="openapi_release_notes/slurm-24.11.4.html"-->
+<!--#include virtual="openapi_release_notes/slurm-24.11.3.html"-->
+<!--#include virtual="openapi_release_notes/slurm-24.11.2.html"-->
+<!--#include virtual="openapi_release_notes/slurm-24.11.1.html"-->
 <!--#include virtual="openapi_release_notes/slurm-24.11.0.html"-->
 <!--#include virtual="openapi_release_notes/slurm-24.05.3.html"-->
 <!--#include virtual="openapi_release_notes/slurm-24.05.1.html"-->
@@ -83,6 +87,6 @@ clients sooner than would be otherwise required.</p>
 
 <hr size=4 width="100%">
 
-<p style="text-align:center;">Last modified 25 August 2023</p>
+<p style="text-align:center;">Last modified 8 April 2025</p>
 
 <!--#include virtual="footer.txt"-->
diff --git a/doc/html/openapi_release_notes/slurm-24.11.1.html b/doc/html/openapi_release_notes/slurm-24.11.1.html
@@ -0,0 +1,29 @@
+<h2 id="24111">Slurm 24.11.1<a class="slurm_link" href="#24111"></a></h2>
+<h3>data_parser/v0.0.42</h3>
+<h4>Modified Fields</h4>
+<table class="tchange">
+<tr class="trchange">
+	<td class="tdfield">Replaced</td>
+	<td class="tdchange">.components.schemas["v0.0.42_job_desc_msg"].properties.argv.description = "Arguments to the script. Note: The slurmstepd always overrides argv[0] with the path to the created script file. If this option is used, argv[0] should be a throwaway value.";</td>
+</tr>
+<tr class="trchange">
+	<td class="tdfield">Replaced</td>
+	<td class="tdchange">.info.version = "Slurm-24.11.1&openapi/slurmdbd&openapi/slurmctld";</td>
+</tr>
+</table>
+<h3>data_parser/v0.0.41</h3>
+<h4>Modified Fields</h4>
+<table class="tchange">
+<tr class="trchange">
+	<td class="tdfield">Replaced</td>
+	<td class="tdchange">.info.version = "Slurm-24.11.1&openapi/slurmdbd&openapi/slurmctld";</td>
+</tr>
+</table>
+<h3>data_parser/v0.0.40</h3>
+<h4>Modified Fields</h4>
+<table class="tchange">
+<tr class="trchange">
+	<td class="tdfield">Replaced</td>
+	<td class="tdchange">.info.version = "Slurm-24.11.1&openapi/slurmdbd&openapi/slurmctld";</td>
+</tr>
+</table>