[debian-mysql] Bug#996028: InnoDB: corrupted TRX_NO after upgrading to 10.3.31

[Marc Gallet]

> With 10.3.29 running, I've dumped all databases (mysqldump --all-databases -p >mysql.dump),
> then dropped all databases, stopped mariadb and deleted /var/lib/mysql/ib*.
> Then restarted mariadb, restored databases (source mysql.dump) and finally upgraded to 10.3.31.
>
> So my problem is solved. I've also found the patch that's causing the problem.
> I'm not going to waste any more time on this.
> 
> During many years of running mysql, upgrade just worked all the time.
> Looks like there are some quality problems with the developement process now.

First of all, I fully understand that maintaining is a voluntary work and that
you are free to do as you please =) In the following, I am not meaning *you*
(or anyone else) "must" do anything they don't want to.
Merely raising some questions.

Am I to understand that the expected path forward with what is supposed to be
a minor update offered on oldstable is that everyone shall dump their databases,
delete the data folder, restore, then upgrade???

Sure I can do that, not happy to, but I can.

However, that's an awful upgrade path and user experience for what would have
normally been a mere and uneventful "apt-get upgrade". I do hope not everyone
still on buster will have to do that after learning about it from a bug report.

Seems to me it would be better to not have pushed this update to oldstable in
the first place. (I'm not familiar with process, but could it be possibly undone,
such that users stay on 10.3.29 and are not proposed the broken upgrade?).

Any chance for a mitigating change coming from Debian's side? (not thinking about
my specific case, but rather the general case of the thousands or more of people
using mariadb on buster currently)

Marc