I think this never happened to me before. How to debug it?
$ rpm-ostree update
error: While pulling fedora/38/x86_64/silverblue: While fetching mirrorlist 'https://ostree.fedoraproject.org/mirrorlist': While fetching https://ostree.fedoraproject.org/mirrorlist: [27] Out of memory
$ free -m
total used free shared buff/cache available
Mem: 7786 995 4682 222 2108 6319
Swap: 7785 0 7785
[fede@fedora ~]$ rpm-ostree update
error: While pulling fedora/38/x86_64/silverblue: While fetching mirrorlist 'https://ostree.fedoraproject.org/mirrorlist': While fetching https://ostree.fedoraproject.org/mirrorlist: [27] Out of memory
Faulty local DNS is a common issue I’ve been dealing for years and it is not limited to this forum.
Often users are not knowledgeable enough to properly investigate the problem, thus addressing their negative experience to Linux in general and Fedora in particular.
This specifically happens when the source of the problem is poor compatibility with EDNS(0), or the difference in DNS query methods between Linux and Windows.
I wonder if dealing with DNS similar to NTP by preferring a global provider over a local one might lead to a better experience at large scale.
Just to be clear: the Internet connection works fine with this “faulty” DNS; only rpm-ostree is failing.
I always use global DNS servers such as 1.1.1.1, but I’ve just relocated and I was testing the connection with a new provider so I changed the router settings DNS to those of the provider.
$ grep nameserver /etc/resolv.conf
nameserver 127.0.0.53
$ rpm-ostree update
error: While pulling fedora/38/x86_64/silverblue: While fetching mirrorlist 'https://ostree.fedoraproject.org/mirrorlist': While fetching https://ostree.fedoraproject.org/mirrorlist: [27] Out of memory
Keep in mind that one domain failing to resolve correctly means there may be others, perhaps not so critical as to be immediately identifiable, but still causing minor glitches or timeouts, which you may not even notice if the service falls back from its optimal operation method to sub-optimal, just to not fail entirely.
Now it’s failing despite the global DNS servers…
I must say that I had to set them in GNOME network settings, as setting them in my Fritz router settings was not enough maybe.
I’ve just realized that I should have typed resolvectl status to be sure which DNS servers have been used.
$ resolvectl status
Global
Protocols: LLMNR=resolve -mDNS -DNSOverTLS DNSSEC=no/unsupported
resolv.conf mode: stub
Link 2 (wlp1s0)
Current Scopes: DNS LLMNR/IPv4 LLMNR/IPv6
Protocols: +DefaultRoute LLMNR=resolve -mDNS -DNSOverTLS DNSSEC=no/unsupported
Current DNS Server: 8.8.8.8
DNS Servers: 8.8.8.8 1.1.1.1 192.168.178.1
DNS Domain: fritz.box
[fede@fedora ~]$ rpm-ostree update
error: While pulling fedora/38/x86_64/silverblue: While fetching mirrorlist 'https://ostree.fedoraproject.org/mirrorlist': While fetching https://ostree.fedoraproject.org/mirrorlist: [27] Out of memory
You can enable DoT to isolate the issue and prevent possible DNS hijacking.
It the problem persists, it is likely caused by IP connectivity and unrelated to DNS.
Perhaps some of the server IPs are unreachable for you for some reason.
It’s happening again despite the good DNS servers. But the RAM is available:
$ rpm-ostree update
error: While pulling fedora/38/x86_64/silverblue: While fetching mirrorlist 'https://ostree.fedoraproject.org/mirrorlist': While fetching https://ostree.fedoraproject.org/mirrorlist: [27] Out of memory
[fede@fedora ~]$ free -m
total used free shared buff/cache available
Mem: 7785 1798 2370 328 3616 5339
Swap: 7784 118 7666
For sure I have DNS problems in my Fedora laptop.
Today my laptop is connected to Internet using my phone hotspot (DNS set by the provider). I can’t access a website on my Fedora laptop, but I can access it in my phone. They use the same connection settings.
Maybe the systemd DNS resolver. I vaguely remember it caused problems in the past.