The Qlustar team is pleased to announce the immediate availability of
Qlustar 13 for download [1]. It updates Qlustar's core platform to
current Ubuntu 22.04 LTS. The AlmaLinux [2] edge platform remains based
on 8.6 while CentOS 7 [3] is still supported via Qlustar 12 images and
the new cross-version compatibility.
The highlights of this release are the addition of BMC hardware
management via IPMI/Redfish in QluMan, the integration of Spack as the
new HPC package manager and the cross-version compatibility between
Qlustar releases.
QluMan's new *BMC hardware management* support allows for BMC
authentification management, power control of nodes via IPMI or Redfish,
the display of BIOS settings as well as their validation and
correction. A cumulative problem report showing an overview of cluster
node misconfigurations has also been added to QluMan.
Spack [4] has been introduced as the new package manager for HPC related
application software beyond the so-called *Qlustar HPC core stack*. It
offers a huge list of supported software packages and by design provides
hardware optimized versions of them.
The HPC core stack is a new component definition within Qlustar and is
related to the introduction of Spack. Rather than aiming to provide as
much HPC functionality as possible by OS packages (debs/rpms), the HPC
core stack now makes a clear cut between Qlustar related system software
which remains provided as OS packages and HPC application software.
Starting with this release, *cross-version compatibility* between the
current and the previous Qlustar release has been added. In practice
this means, that on a cluster with a Qlustar 13 head-node you can still
run nodes with Qlustar 12 images/chroots and vice versa. This new
feature was added to allow even smoother release upgrades.
Highlights among the various *major component updates* include Kernel
5.15.x, Slurm 22.05.x, Lustre 2.15.1, ZFS 2.1.6 and BeeGFS 7.3.2. Please
read the release notes [5] for more details.
[1] https://qlustar.com/download/
[2] https://almalinux.org/
[3] https://www.centos.org/
[4] https://spack.readthedocs.io/en/latest/
[5] https://docs.qlustar.com/Qlustar/13/ClusterOS/release-notes.html
The Qlustar releases 12.0.2.0-b556f1460 and 11.0.1.19-b549f1462 are ready
including a number of security/bug fixes and improvements. Please check
the following web pages for details about security fixes and special
update instructions:
https://qlustar.com/qsa/2022/1219221https://qlustar.com/qsa/2022/1219222
The following non-security related updates/enhancements/bug fixes are included:
For 12.0 only:
- Update to QluMan 13.0.2 (check the upcoming Qlustar 13 release notes
for a description of new features)
- Update Nvidia drivers to 515.86.01
- Update to pmix 4.1.2
- Update to slurm 22.05.6
- Update to ucx 1.13.0
- Update to ZFS 2.1.6
The Qlustar releases 12.0.1.1-b548f1445 and 11.0.1.18-b549f1444 are ready
including a number of security/bug fixes and improvements. Please check
the following web pages for details about security fixes and special
update instructions:
https://qlustar.com/qsa/2022/1029221https://qlustar.com/qsa/2022/1029222
The Qlustar releases 12.0.1.0-b548f1441 and 11.0.1.17-b549f1440 are ready
including a number of security/bug fixes and improvements. Please check
the following web pages for details about security fixes and special
update instructions:
https://qlustar.com/qsa/2022/0808221https://qlustar.com/qsa/2022/0808222
The following non-security related enhancements/bug fixes are included:
For 12.0 only:
- Update to QluMan 12.0.7.5
- Update Nvidia drivers to 515.65.01 (fixes CVE-2022-31607,
CVE-2022-31608, CVE-2022-31615)
- Update to ZFS 2.1.5
- Added Lustre 2.15
A number of new Qlustar features were completed recently and published
together with the latest security updates.
QluMan 12.0.7.3 highly improves on its slurm management capabilities
basically making the slurm component feature complete. It adds
- Real SLURM QOS MANAGEMENT [1] in the GUI. QoS can now easily be added,
edited, removed and assigned. The full list of QoS properties may be
added/removed and manipulated including sanity checking the values
entered.
- SLURM ASSOCIATION AND LIMITS MANAGEMENT [2] has been added in the GUI
unleashing the power of advanced slurm resource management to QluMan
admins. Various visual filters have been created to achieve a
comprehensive display of slurm associations. The full list of
association settings may be added/removed and manipulated including
sanity checking the values entered.
* An 'LDAP GROUP <-> SLURM ACCOUNT' SYNC MECHANISM [3] has been designed
to automate slurm user setup upon creation of the user in LDAP. Slurm
accounts may be associated with LDAP groups such that depending on
users primary group the corresponding slurm account will be assigned
as their default.
* Cron-based LDAP DB BACKUP was introduced to allow a disaster recovery
in case the LDAP DB becomes corrupted.
Other new features:
- Update to AlmaLinux 8.6
[1] https://docs.qlustar.com/Qlustar/12.0/ClusterOS/qluman-guide/components/slu…
[2] https://docs.qlustar.com/Qlustar/12.0/ClusterOS/qluman-guide/components/slu…
[3] https://docs.qlustar.com/Qlustar/12.0/ClusterOS/qluman-guide/components/slu…
The Qlustar releases 12.0.0.12-b548f1436 and 11.0.1.16-b549f1437 are ready
including a number of security/bug fixes and improvements. Please check
the following web pages for details about security fixes and special
update instructions:
https://qlustar.com/qsa/2022/0624221https://qlustar.com/qsa/2022/0624222
The following non-security related enhancements/bug fixes are included:
For 12.0 only:
- Update to AlmaLinux 8.6
- Update to QluMan 12.0.7.3
- Update to MLNX OFED 5.4-3.4.0.0
- Update to Lustre 2.12.9
The Qlustar releases 12.0.0.11-b547f1433 and 11.0.1.15-b543f1435 are ready
including a number of security/bug fixes and improvements. Please check
the following web pages for details about security fixes and special
update instructions:
https://qlustar.com/qsa/2022/0524221https://qlustar.com/qsa/2022/0524222https://qlustar.com/qsa/2022/0524223
The following non-security related enhancements/bug fixes are included:
For 12.0 only:
- Update to QluMan 12.0.7.1
- Update to slurm 21.08.8.2
The Qlustar releases 12.0.0.10-b547f1432 and 11.0.1.14-b543f1431 are ready
including a number of security/bug fixes and improvements. Please check
the following web pages for details about security fixes and special
update instructions:
https://qlustar.com/qsa/2022/0413221https://qlustar.com/qsa/2022/0413222
The following non-security related enhancements/bug fixes are included:
For 12.0 only:
- Update to Nvidia graphics driver version 510.47.03
A number of new Qlustar features were completed recently and published
together with the latest security updates.
QluMan 12.0.6.0 has the following new capabilities:
- QLUMAN TIMERS [1] present a new Config class assignable to nodes. They
allow for time-based execution of commands on nodes and are easily
configured via the GUI.
- ENHANCEMENTS OF THE SLURM GUI COMPONENT [2]:
* Added on-the-fly job filtering and job grouping.
* Significantly improved mechanism to transfer job data. The QluMan
GUI is now able to smoothly handle thousands of jobs.
- HARDWARE DETECTION INFO for nodes in the GUI Enclosure view [3] now
transparently highlights misconfigurations, helping admins to track
down faulty settings.
- FILTERS IN USER/GROUP MANAGEMENT [4] have been added, making LDAP management
a lot clearer and easier to follow.
Other new features:
- MELLANOX OFED [5] drivers and tools have been included for optimized support
of clusters with newest Infiniband (IB) technology. Qlustar now
automatically detects whether IB adapters supported by the current
Mellanox OFED release are present and activates its drivers if that's
the case. For all other IB hardware, the in-kernel drivers are used.
- GPU DIRECT RDMA [6] support has been added for nodes running Nvidia GPUs
together with IB adapters supported by Mellanox OFED. It provides a
significant decrease in GPU-GPU communication latency and completely
offloads the CPU, removing it from all GPU-GPU communications across
the network.
- The VIRTUALGL image module is now production ready. When used in a
node image, Qlustar can provide 3D remote visualization on nodes with
the required hardware.
[1] https://docs.qlustar.com/Qlustar/12.0/ClusterOS/qluman-guide/Config-Classes…
[2] https://docs.qlustar.com/Qlustar/12.0/ClusterOS/qluman-guide/components/slu…
[3] https://docs.qlustar.com/Qlustar/12.0/ClusterOS/qluman-guide/Adding-Hosts.h…
[4] https://docs.qlustar.com/Qlustar/12.0/ClusterOS/qluman-guide/components/lda…
[5] https://www.mellanox.com/products/infiniband-drivers/linux/mlnx_ofed
[6] https://www.mellanox.com/products/GPUDirect-RDMA