Actually on debian we use the standardly available sysfsutils, which I recommend for this (and which is available in Redhat, too) ; it is the equivalent for /sys than sysctl is for /proc:
1) install sysfsutils either with apt install or dnf install 2) config the stuff: echo "kernel/mm/transparent_hugepage/enabled = madvise" > /etc/sysfs.d/transparent_hugepage.conf echo "kernel/mm/transparent_hugepage/defrag = madvise" >> /etc/sysfs.d/transparent_hugepage.conf systemctl enable sysfsutils.service 3) either reboot or just restart the services including unbound: systemctl restart sysfsutils.service systemctl restart unbound.service That's it, the conf will survive reboot and updates. Le sam. 16 mars 2024 à 14:08, Yuri via Unbound-users < unbound-users@lists.nlnetlabs.nl> a écrit : > It is necessary to disable THP so often that I had to write a script for > this purpose. Doing this using the service is much more convenient than > doing it any other way, in my opinion. > > https://github.com/yvoinov/memory-tools/blob/main/disable_thp.sh > 16.03.2024 14:23, Tarko Tikan via Unbound-users пишет: > > hey, > > The question is that stats metrics (mem*) are stable . They rapidly grow > after startup and stay at a logical > maximum and don't keep growing. > > But the process size does. > > > We saw the same on Debian due to THP like Oliver already mentioned. See > https://github.com/NLnetLabs/unbound/issues/724 > > We "solved" this by disabling THP on our unbound machines. > > -- *Ce message et toutes les pièces jointes (ci-après le "message") sont établis à l’intention exclusive des destinataires désignés. Il contient des informations confidentielles et pouvant être protégé par le secret professionnel. Si vous recevez ce message par erreur, merci d'en avertir immédiatement l'expéditeur et de détruire le message. Toute utilisation de ce message non conforme à sa destination, toute diffusion ou toute publication, totale ou partielle, est interdite, sauf autorisation expresse de l'émetteur*
