Hi, I ran a sudo dnf update which updated and loaded the two F41 Updates repositories and then put on the updates. Immediately after the update finished I ran a dnf needs-restarting which said there was no reboot required, but before it said that it updated and loaded all the existing repositories including the F41 Updates repository, why was that updated again, surely that repository was flagged as up to date by the prior dnf update?
Having said this though a subsequent dnf needs-restarting -s didn't update any of the repositories but did list a whole host of services that needed restarting, which included sddm and akmods, which I didn't understand because as far as I could see there was nothing updated by dnf that was related to sddm nor was there anything related to akmods or nvidia drivers.
regards, Steve
On Thu, 2024-11-07 at 09:11 +1100, Stephen Morris wrote:
Hi, I ran a sudo dnf update which updated and loaded the two F41 Updates repositories and then put on the updates. Immediately after the update finished I ran a dnf needs-restarting which said there was no reboot required, but before it said that it updated and loaded all the existing repositories including the F41 Updates repository, why was that updated again, surely that repository was flagged as up to date by the prior dnf update?
Having said this though a subsequent dnf needs-restarting -s didn't update any of the repositories but did list a whole host of services that needed restarting, which included sddm and akmods, which I didn't understand because as far as I could see there was nothing updated by dnf that was related to sddm nor was there anything related to akmods or nvidia drivers.
(This may or may not be relevant to the needs-restarting issue)
I've posted a BZ report about the problem with tracer:
https://bugzilla.redhat.com/show_bug.cgi?id=2324183
poc
On Wed, Nov 6, 2024 at 5:26 PM Patrick O'Callaghan pocallaghan@gmail.com wrote:
On Thu, 2024-11-07 at 09:11 +1100, Stephen Morris wrote:
Hi, I ran a sudo dnf update which updated and loaded the two F41 Updates repositories and then put on the updates. Immediately after the update finished I ran a dnf needs-restarting which said there was no reboot required, but before it said that it updated and loaded all the existing repositories including the F41 Updates repository, why was that updated again, surely that repository was flagged as up to date by the prior dnf update?
Having said this though a subsequent dnf needs-restarting -s didn't update any of the repositories but did list a whole host of services that needed restarting, which included sddm and akmods, which I didn't understand because as far as I could see there was nothing updated by dnf that was related to sddm nor was there anything related to akmods or nvidia drivers.
(This may or may not be relevant to the needs-restarting issue)
I've posted a BZ report about the problem with tracer:
Does the tip of Master work?
git clone https://github.com/FrostyX/tracer.git sudo tracer/bin/tracer.py -ea
If so, it might make a good datapoint for your BZ.
https://bugzilla.redhat.com/show_bug.cgi?id=1214961 shows some other troubleshooting steps.
Jeff
On 6 Nov 2024, at 22:11, Stephen Morris steve.morris.au@gmail.com wrote:
Hi, I ran a sudo dnf update which updated and loaded the two F41 Updates repositories and then put on the updates. Immediately after the update finished I ran a dnf needs-restarting which said there was no reboot required, but before it said that it updated and loaded all the existing repositories including the F41 Updates repository, why was that updated again, surely that repository was flagged as up to date by the prior dnf update?
There are two things going on. 1. The update repo is being continuously updated. 2. The time when you download the repo meta data.
I always use `dnf update --refresh` to ensure that the meta data is up to date. But if you run the command again new packages may have been released.
Having said this though a subsequent dnf needs-restarting -s didn't update any of the repositories
It is not it’s job to update repositories. It’s job is tell you if services are using out of date code.
but did list a whole host of services that needed restarting, which included sddm and akmods, which I didn't understand because as far as I could see there was nothing updated by dnf that was related to sddm nor was there anything related to akmods or nvidia drivers.
The services maybe using a library that is common, for example glibc.
Barry
regards, Steve
<OpenPGP_0x1EBE7C07B0F7242C.asc>
users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
On Thu, 2024-11-07 at 01:54 -0500, Jeffrey Walton wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=1214961 shows some other troubleshooting steps.
That report is old but interesting. I tried:
$ sudo tracer --show-resource=packages Date Time Package name ------------------------------------------------------- $
For some reason it can't find a list of packages, which would explain a lot. The we have:
$ ls -l /var/lib/dnf/history.sqlite -rw-r--r--. 1 root root 2646016 Oct 31 09:29 /var/lib/dnf/history.sqlite
i.e. the 'history' database may be out of date. Nevertheless, 'dnf5 history' shows updates to today, so maybe not.
poc
On Nov 6, 2024, at 17:11, Stephen Morris steve.morris.au@gmail.com wrote:
why was that updated again, surely that repository was flagged as up to date by the prior dnf update?
Not 100% sure but I suspect something in dnf needs-restarting is looking for the extra dnf metadata that includes package file lists, which aren’t downloaded for normal updates, to speed up regular updates (since they’re significantly larger metadata files).
Just like when you run “dnf install httpd” it only downloads the smaller set of information, but if you then run “dnf install /usr/bin/top” it has to download more metadata, even though you just updated a moment ago.
On 7/11/24 19:30, Barry wrote:
On 6 Nov 2024, at 22:11, Stephen Morrissteve.morris.au@gmail.com wrote:
Hi, I ran a sudo dnf update which updated and loaded the two F41 Updates repositories and then put on the updates. Immediately after the update finished I ran a dnf needs-restarting which said there was no reboot required, but before it said that it updated and loaded all the existing repositories including the F41 Updates repository, why was that updated again, surely that repository was flagged as up to date by the prior dnf update?
There are two things going on.
- The update repo is being continuously updated.
- The time when you download the repo meta data.
I always use `dnf update --refresh` to ensure that the meta data is up to date. But if you run the command again new packages may have been released.
Having said this though a subsequent dnf needs-restarting -s didn't update any of the repositories
It is not it’s job to update repositories. It’s job is tell you if services are using out of date code.
but did list a whole host of services that needed restarting, which included sddm and akmods, which I didn't understand because as far as I could see there was nothing updated by dnf that was related to sddm nor was there anything related to akmods or nvidia drivers.
The services maybe using a library that is common, for example glibc.
Thanks Barry, that might explain it. I did notice with one of the packages installed there was a glibc scriptlet run, which interestingly didn't display the deprecated spawning warning.
regards, Steve
Barry
regards, Steve
<OpenPGP_0x1EBE7C07B0F7242C.asc>
users mailing list --users@lists.fedoraproject.org To unsubscribe send an email tousers-leave@lists.fedoraproject.org Fedora Code of Conduct:https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines:https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives:https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org Do not reply to spam, report it:https://pagure.io/fedora-infrastructure/new_issue
On 8/11/24 08:57, Jonathan Billings wrote:
On Nov 6, 2024, at 17:11, Stephen Morrissteve.morris.au@gmail.com wrote:
why was that updated again, surely that repository was flagged as up to date by the prior dnf update?
Not 100% sure but I suspect something in dnf needs-restarting is looking for the extra dnf metadata that includes package file lists, which aren’t downloaded for normal updates, to speed up regular updates (since they’re significantly larger metadata files).
Just like when you run “dnf install httpd” it only downloads the smaller set of information, but if you then run “dnf install /usr/bin/top” it has to download more metadata, even though you just updated a moment ago.
Thanks Jonathan, I had forgotten about that issue that causes problems with dnfdragora.
regards, Steve
On Thu, Nov 7, 2024 at 4:57 PM Jonathan Billings billings@negate.org wrote:
On Nov 6, 2024, at 17:11, Stephen Morris steve.morris.au@gmail.com wrote:
why was that updated again, surely that repository was flagged as up to date by the prior dnf update?
Not 100% sure but I suspect something in dnf needs-restarting is looking for the extra dnf metadata that includes package file lists, which aren’t downloaded for normal updates, to speed up regular updates (since they’re significantly larger metadata files).
Just like when you run “dnf install httpd” it only downloads the smaller set of information, but if you then run “dnf install /usr/bin/top” it has to download more metadata, even though you just updated a moment ago.
If stale metadata is the case, then rebuilding the dnf cache with the --refresh option may sidestep the problem. Something like `sudo dnf makecache --refresh`. Also see https://dnf.readthedocs.io/en/latest/command_ref.html#metadata-synchronization.
Jeff
On Thu, Nov 7, 2024 at 4:57 PM Stephen Morris steve.morris.au@gmail.com wrote:
On 7/11/24 19:30, Barry wrote:
On 6 Nov 2024, at 22:11, Stephen Morris steve.morris.au@gmail.com wrote:
I ran a sudo dnf update which updated and loaded the two F41 Updates repositories and then put on the updates. Immediately after the update finished I ran a dnf needs-restarting which said there was no reboot required, but before it said that it updated and loaded all the existing repositories including the F41 Updates repository, why was that updated again, surely that repository was flagged as up to date by the prior dnf update?There are two things going on.
- The update repo is being continuously updated.
- The time when you download the repo meta data.
I always use `dnf update --refresh` to ensure that the meta data is up to date. But if you run the command again new packages may have been released.
Having said this though a subsequent dnf needs-restarting -s didn't update any of the repositories
It is not it’s job to update repositories. It’s job is tell you if services are using out of date code.
but did list a whole host of services that needed restarting, which included sddm and akmods, which I didn't understand because as far as I could see there was nothing updated by dnf that was related to sddm nor was there anything related to akmods or nvidia drivers.
The services maybe using a library that is common, for example glibc.
Thanks Barry, that might explain it. I did notice with one of the packages installed there was a glibc scriptlet run, which interestingly didn't display the deprecated spawning warning.
The noisy glibc trigger-post-uninstall scriptlet was fixed recently; see https://bugzilla.redhat.com/show_bug.cgi?id=2291869.
Jeff
On 8/11/24 09:08, Jeffrey Walton wrote:
The noisy glibc trigger-post-uninstall scriptlet was fixed recently; seehttps://bugzilla.redhat.com/show_bug.cgi?id=2291869.
Thanks Jeffrey, my assumption was it had been updated.
regards, Steve
-
Barry -
Jeffrey Walton -
Jonathan Billings -
Patrick O'Callaghan -
Stephen Morris