Ask Your Question

Alan Ian Ross's profile - activity

2018-11-18 20:59:47 -0500 received badge  Necromancer (source)
2018-11-18 20:59:47 -0500 received badge  Teacher (source)
2017-08-03 11:16:06 -0500 answered a question Failed to synchronize cache for repo 'updates' fedora25

Unfortunately, the one described in the opening question is still the standard behavior in Fedora 26.

Edition: Fedora 26 Server (no graphical user interface, no windows manager, only the barebones system) installed as a virtual system in VMware WS Player 12.5.7 running on top of Ubuntu 16.04.3.

Behavior: the

# dnf update

command still gets

Error: Failed to synchronize cache for repo 'updates'

when used

  1. with networking set up in bridged mode in VMware, and
  2. repos (both in /etc/yum.repos.d/fedora.repo and in /etc/yum.repos.d/fedora-updates.repo) address format set up in https, wherever it is set up as such in the two config files.

Turning "https" into "http" makes the error disappear, and dnf returns to its regular behavior. Unfortunate: but true.

The issue, by the way, can't be related with some mistake in the network setup, or a malfunctioning in my internet connection: because, for example,

$ lynx

(or, for that matter, any other URL) works perfectly.

By the way n°2: if, instead of using the VMware network infrastructure in bridged mode, I use it in NAT mode, dnf keeps on working flawlessly, even without turning the "https" into "http" in the two previously mentioned config files...

N.B.: none of the other fixes found around (here and elsewhere), from adding a


line in /etc/dnf/dnf.conf (see e.g. here), to reinstalling ca-certificates-2017.2.14-1.0.fc26.noarch.rpm (e.g. here), to (re)installing "librepo / python-librepo" (e.g. here), seems to work in this case.

For the record, I add that I've recently upgraded from Fedora 25 to Fedora 26 (with the same underlying VMware setup): but in Fedora 25 the very same behavior happened.

Hope this can help (maybe for a diagnosis)...