The Qlustar releases 12.0.0.2-b531f1351 and 11.0.1.6-b533f1350 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/news/qsa-0320211-linux-kernel-vulnerabilities https://qlustar.com/news/qsa-0320212-security-update-bundle
The following non-security related enhancements/bug fixes are included:
For 12.0 only: - QluMan 12.0.1.0 * Add support for home directories on servers other than the head-node * Network FS mounts: respect Network Override * Improve saving of window geometries * Allow cluster commands to execute on headnodes and show correct execute-on in ExecuteWidget * Fix exception when importing a newly generated auth token * NetworkConfigWidget: Adjust column sizes * Nameservices: add group membership to passwd-sync method * Nameservices: fix permission of sssd.conf on updates * RxEngine: Fix kill/redo button in tabs
- Set Infiniband node_desc to the node's hostname on nodes with IB adapters. This allows for better IB fabric analysis.
- Prevent motd to display twice upon login on CentOS 8.
Hi Roland,
Thank you very much for the updates, glad to see my little permissions patch made it in!
I've got another potential one for you - I've noticed that even after the new updates, and even when defining Network Mounts for <headnode>:/srv/apps and <headnode>:/srv/data as /apps and /data for compute nodes, they don't automount for some reason. When I manually mount those on each node:
mount -t nfs <headnode>:/srv/data /data mount -t nfs <headnode>:/srv/apps /apps
they work fine and persist until the next reboot. I'm not sure which file(s) control pushing the network mounts to the compute nodes, though I do notice that the <headnode>:/srv/ql-common and <headnode>:/srv/apps/chroots/focal do mount correctly.
Any idea what might be causing this? I've reproduced it on both my real, legit cluster with identical, rack-mount hardware and my cobbled-together laptop cluster.
Thanks, -Mike
"M" == hereiam hereiam@mit.edu writes:
Hi Mike,
M> Hi Roland, Thank you very much for the updates, glad to see my M> little permissions patch made it in!
yup, thanks again for reporting.
M> I've got another potential one for you - I've noticed that even M> after the new updates, and even when defining Network Mounts for M> <headnode>:/srv/apps and <headnode>:/srv/data as /apps and /data M> for compute nodes, they don't automount for some reason. When I M> manually mount those on each node:
>>>> mount -t nfs <headnode>:/srv/data /data mount -t nfs >>>> <headnode>:/srv/apps /apps
M> they work fine and persist until the next reboot. I'm not sure M> which file(s) control pushing the network mounts to the compute M> nodes, though I do notice that the <headnode>:/srv/ql-common and M> <headnode>:/srv/apps/chroots/focal do mount correctly.
M> Any idea what might be causing this? I've reproduced it on both M> my real, legit cluster with identical, rack-mount hardware and my M> cobbled-together laptop cluster.
We can't reproduce this. Note that newly defined mounts are available only after a reboot of the nodes and only after accessing the mount point (automount). You can use the context-menu entry 'Preview config' when on a host in the Enclosure View to see the generated files for this node when it boots.
Best,
Roland