September 2024 - AlmaLinux Users - AlmaLinux List Archives

Autofs: too many open files issue
by Carles Acosta Silva 24 Nov '24

24 Nov '24

Hi, We are running AlmaLinux 9.3 on several nodes that normally use autofs to mount different kinds of remote directories. Since the last week, we are seeing these errors: automount[2150866]: open_pipe:161: failed to open pipe: Too many open files And automount stops working. It does not seem that any user has arrived at the maximum number of open files. Do you have any idea where we can look to find a solution? Thank you in advance. Best regards, Carles

4 4

AlmaLinux 9.3 and libvirtd problem
by Alessandro Baggi 22 Oct '24

22 Oct '24

Hi list, I'm running AlmaLinux 9.3 and I'm running several virtual machines using qemu-kvm and libvirtd ( and virt-manager). 3 VMs starts during system boot (configured with "virsh autostart VMNAME"). All works as expected until I loaded a old vm config file. When I tried to run "systemctl reload libvirtd" I got: "libvirtd.service is not active, cannot reload." It is enable on boot and running "systemctl status libvirtd" I get: Loaded: loaded (/usr/lib/systemd/system/libvirtd.service; enabled; preset: disabled) Active: inactive (dead) since Thu 2024-01-25 16:32:40 CET; 4min 55s ago Duration: 5min 14.530s It reports that it is inactive (really dead) because I can't find it inside process table. Rebooting the system, I found libvirtd process with argument "--timeout=120" and after 120 seconds it dies. Reading libvirtd systemd unit I found Environment=LIBVIRTD_ARGS="--timeout 120" and from the man page I get: "t, --timeout *SECONDS* Exit after timeout period (in seconds), provided there are neither any client connections nor any running domains." Here I found something strange. On boot I have 3 running domains but libvirtd dies after 120 seconds. To load new configuration I need to run restart and virt-manager works without problem also if libvirtd is not running. This is a misconfiguration or it is a bug? Thank you in advance. Alessandro.

2 4

Kickstart installs with isolinux (insecure boot) fails, same Kickstart works with EFI(secureboot)
by fait＠uga.edu 10 Sep '24

10 Sep '24

I have been working with AlmaLinux 8 for a couple of years using various iterations AlmaLinux 8, with my systems(5 of them) currently using 8.10. I am now in the process of validating install media to move to version 9.4, using kickstart to manage the initial configuration of these hosts. The kickstart file sets up the network,decides the host name, and sets up a complex partitioning scenario. Initially I tested this using secure boot, without any issue. I repeated the same install, on the same hardware, using the same kickstart file, only this time using the legacy boot (insecure). I have verified that this is an issue with insecure boot, as I then did a manual install of a server+graphics, along with development tools, and then took the resulting kickstart file, made the required changes so that the partitioning would work, and then attempted to do a kickstart install using this minimal kickstart, with the same results of success for the EFI boot, and failure with the isolinux boot. I did check, and the two commands to the kernel appear to be identical, but the insecure boot is not able to find the kickstart file. Has anyone else seen this behavior? Anyone found a work-around?

1 0

Almalinux 9.4 import LVM VG bug?
by Alessandro Baggi 08 Sep '24

08 Sep '24

Hi list, on my Almalinux 9.4 I tried to import LVM Volume Group from another system. When I attached disks on my workstation, pvscan or vgimport does not found any lvm devices. Reading here: https://discussion.fedoraproject.org/t/lvm-operations-might-fail-on-a-fedor… seems that renaming /etc/lvm/devices/system.devices permits the system to read lvm devices and able to import an existing volume. This is a bug or a wanted behaviour? File system.devices is required? I should restore it? Thank you in advance Alessandro.

2 2