Hi!
I did a bunch of reproducible experiments using Salsa-CI in
https://salsa.debian.org/mariadb-team/mariadb-server/-/pipelines/536587
testing:
## upgrade to Bookworm
* cacti and Bullseye upgrade
- apt install -qq --yes cacti
-> - apt full-upgrade -qq --yes
* default-mysql-server and
Yes, systemd was installed. It was a Debian installed from scratch in
version 11 from the available Linux templates of my cloud provider.
I have installed Apache, MariaDB, PHP which is required for vBulletin
(installed manually, no package). And some other tools which should not
have an
Can you attach the full log as an attachment?
The current short log does not show what apt told you in the beginning
about what it plans to do, nor can I see if mariadb-client-10.5 was
uninstalled or what happened.
But it did at least hit a bug where the uninstall fails on
`invoke-rd.c stop
Le 09/06/2023 à 16:17, Otto Kekäläinen a écrit :
Thanks Frederick for the report!
Do you still have the output of apt? Can you copy-paste here or attach
log to show exactly what happened?
It is likely that you hit this bug, but exact details would help
confirm, and also help build CI
Thanks Frederick for the report!
Do you still have the output of apt? Can you copy-paste here or attach log
to show exactly what happened?
It is likely that you hit this bug, but exact details would help confirm,
and also help build CI test/scenario to ensure we have realistic upgrade
testing.
Hi,
I have just upgraded Debian from 11 to 12, and its mariadb-server. The
command was simply "apt full-upgrade".
I have get the #1035949 bug even if I havn't the zoph package, nor
default-mysql-server.
I didn't manage to make it works with:
# apt --fix-broken install mariadb-client
Forwarded: https://jira.mariadb.org/browse/MDEV-28640
For reference:
* The upgrade scenario this MR fixed:
https://salsa.debian.org/mariadb-team/mariadb-server/-/commit/15cbf6e691827608636e6ff7f0a50432f50d0c4f
* Release notes mention:
On 04/06/2023 07.03, Otto Kekäläinen wrote:
What do you Andreas suggest we do now?
I'd suggest uploading it to experimental immediately (for NEW
processing) and filing a pre-approval bug and let the release team
decide what to do. This fix will probably be delayed to the first point
Indeed the transitional mariadb-server-10.5 fixes the issue.
What do you Andreas suggest we do now?
It is already past freeze for Bookworm, and this is not just a small
fix but also introduces a new package (albeit transitional). Let me
know how you want to proceed and I can immediately tomorrow
I adjusted your patch a bit as it didn't apply cleanly and pushed it
to https://salsa.debian.org/mariadb-team/mariadb-server/-/merge_requests/47
to replace the transitional mariadb-client-10.5 I had earlier.
Thanks for diving deep in piuparts testing for MariaDB 10.11 and for the patch!
Ideally
Control: tag -1 patch
On 01/06/2023 12.31, Andreas Beckmann wrote:
That worked for me for zoph, but a new failure occurred in tt-rss.
I'm now trying a transitional mariadb-server-10.5 package which seems to
work for both these cases. Will do a complete set of piuparts upgrade
tests for all
Processing control commands:
> tag -1 patch
Bug #1035949 [src:mariadb] mariadb: upgrade issue: mariadb-server-10.5 fails to
stop after all other -10.5 packages were removed
Added tag(s) patch.
--
1035949: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1035949
Debian Bug Tracking Sys
On 26/05/2023 19.16, Otto Kekäläinen wrote:
I filed now
https://salsa.debian.org/mariadb-team/mariadb-server/-/merge_requests/47
as an exploration to fix this issue.
That worked for me for zoph, but a new failure occurred in tt-rss.
I'm now trying a transitional mariadb-server-10.5 package
I filed now
https://salsa.debian.org/mariadb-team/mariadb-server/-/merge_requests/47
as an exploration to fix this issue.
If we don't fix this in 10.11 the alternative would be to patch 10.5
and 10.3 to simply never fail on missing mariadb-client-10.3/5
package. I already did
I am experimenting the suggestion in
https://salsa.debian.org/otto/mariadb-server/-/commits/bugfix/1035949-upgrade-removes-client
On 18/05/2023 05.17, Otto Kekäläinen wrote:
I am not sure what to debug next.
You could try to reintroduce mariadb-client-10.5 as a transitional
package depending on mariadb-client (>= 1:10.11) (and make existing
Conflicts/Breaks/Replaces against mariadb-client-10.5 versioned as (<<
Here is the apt resolver output for debugging:
# apt install default-mysql-server -o Debug::pkgDepCache::Marker=1 -o
Debug::pkgDepCache::AutoInstall=1 -o Debug::pkgProblemResolver=1
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
MarkInstall
I did some more testing in throwaway containers.
In each test starting point was same:
apt-get install default-mysql-server zoph
sed s/bullseye/bookworm/g -i /etc/apt/sources.list
apt update
All cases ran apt 2.6.0.
I only varied the command that followed:
1) apt-get install
On 15/05/2023 02.38, Otto Kekäläinen wrote:
This is not a fix, but could help make the situation easier to
detect/debug for users:
I haven't tried that and have no time to do it right now: if you first
upgrade apt to the bookworm version, and do the dist-upgrade afterwards,
does that change
This is not a fix, but could help make the situation easier to
detect/debug for users:
* https://salsa.debian.org/mariadb-team/mariadb-10.5/-/merge_requests/14
* https://salsa.debian.org/mariadb-team/mariadb-10.3/-/merge_requests/37
Hi Andreas!
Thanks for reporting and looking into this.
> Here apt choses a suboptimal removal order: mariadb-server-10.5 gets
> removed (and therefore stopped, but that fails) only after
> mariadb-client-10.5 and mariadb-client-core-10.5 are already gone.
You are right. The /usr/bin/mysqladmin
Source: mariadb
Version: 1:10.11.2-1
Severity: serious
User: debian...@lists.debian.org
Usertags: piuparts
Control: affects -1 + zoph
Hi,
during a test with piuparts I noticed your package fails to upgrade from
'bullseye'.
It installed fine in 'bullseye', then the upgrade to 'bookworm' fails.
22 matches
Mail list logo
