This is a weird behavior that is happening basically everytime I launch dnf update, which is once every few days. On the first run, it just says dependencies are resolved and there's nothing to do. Just a few seconds later it does find updates. Anybody has any clue as to why this is happening? Screenshot here.
That is what --refresh is for
dnf update --refresh
or
dnf upgrade --refresh
To check for updates without actually upgrading
dnf check-update --refresh
Hi, thank you. Let's say I'm a new user, how do I know I need to do a --refresh if it doesn't find any update on the first run? Please I need you to read my answer to @florian. Your commands do work, but I've only run the second dnf update after a few seconds (see my screenshot in the question), how come that it refreshes metadata automatically only then without me adding any --refresh option?
To know the exact details of how this works, I would need to read the source codes.
@vmercuri Good question, you shouldn't have to run dnf update in order to receive your updates. Generally speaking, the configuration file of your repositories (in /etc/yum.repos.d/) define when it's time to refresh metadata. For example, take a look at any .repo file /etc/yum.repos.d/, you will see a line like
metadata_expire=6h
Meaning: if the metadata stored on your computer is older than 6 hours, dnf will refresh metadata and then determine if there are updates.
Please post /etc/yum.repos.d/fedora-updates.repoto fpaste.org
A fix for Bug 1648274 - dnf fails to refresh expired metadata just came in this morning.
Hi, thank you for your replies. @florian I had already changed the *.repo files and the dnf.conf by disabling the gpgcheck (just to reduce some overhead because dnf is running so slowly on my machine) and setting metadata_expire to 1s (https://paste.fedoraproject.org/paste...). Maybe it's overkill, but I wanted dnf to update metadata at each run, and now it works. @villykruse that's great!
The reason is that your metadata has not yet been refreshed. On your screenshot, you see that they are 1day and 3hours old when you ran dnf update first.
Thanks. I wonder why am I supposed to run that command twice anyway? Shouldn't dnf detect new updates once I run it? What causes the second run to refresh metadata? I've also found this post from another user who basically had the same issue, except I don't have any dnfdragora daemon running in the background.
Do you have dnf-makecache.timer active?
systemctl status dnf-makecache.timer
Thanks. Yes, I do:
dnf-makecache.timer - dnf makecache --timer
Loaded: loaded (/usr/lib/systemd/system/dnf-makecache.timer; enabled; vendor preset: enabled)
Active: active (waiting) since Thu 2019-02-21 17:28:10 CET; 26min ago
Trigger: Thu 2019-02-21 18:38:31 CET; 43min left
Feb 21 17:28:10 fedora systemd[1]: Started dnf makecache --timer.
Asked: 2019-02-15 12:14:06 -0600
Seen: 101 times
Last updated: Feb 16 '19
Ask Fedora is community maintained and Red Hat or Fedora Project is not responsible for content. Content on this site is licensed under a CC-BY-SA 3.0 license.