Debian Bug report logs - #921128
mailman3-web fails to initialize mysql: Specified key was too long

Display info messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Antoine Beaupre <anarcat@debian.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: mailman3-web fails to initialize mysql: Specified key was too long

Date: Fri, 01 Feb 2019 17:52:55 -0500

Package: mailman3-web
Version: 0+20180916-2~bpo9+1
Severity: grave

I can't seem to install mailman3-web, at least from backports:

Paramétrage de mailman3-web (0+20180916-2~bpo9+1) ...
Determining localhost credentials from /etc/mysql/debian.cnf: succeeded.
dbconfig-common: writing config to /etc/dbconfig-common/mailman3-web.conf
mailman3web already exists and has privileges on mailman3web.
creating database mailman3web: already exists.
dbconfig-common: flushing administrative password
Traceback (most recent call last):
  File "/usr/lib/python3/dist-packages/django/db/backends/utils.py", line 64, in execute
    return self.cursor.execute(sql, params)
  File "/usr/lib/python3/dist-packages/django/db/backends/mysql/base.py", line 101, in execute
    return self.cursor.execute(query, args)
  File "/usr/lib/python3/dist-packages/MySQLdb/cursors.py", line 226, in execute
    self.errorhandler(self, exc, value)
  File "/usr/lib/python3/dist-packages/MySQLdb/connections.py", line 36, in defaulterrorhandler
    raise errorvalue
  File "/usr/lib/python3/dist-packages/MySQLdb/cursors.py", line 217, in execute
    res = self._query(query)
  File "/usr/lib/python3/dist-packages/MySQLdb/cursors.py", line 378, in _query
    rowcount = self._do_query(q)
  File "/usr/lib/python3/dist-packages/MySQLdb/cursors.py", line 341, in _do_query
    db.query(q)
  File "/usr/lib/python3/dist-packages/MySQLdb/connections.py", line 280, in query
    _mysql.connection.query(self, query)
_mysql_exceptions.OperationalError: (1071, 'Specified key was too long; max key length is 767 bytes')

That's after stumbling upon bug #919145, so I just had installed mysqldb and
pymysql. I tried purging (deleting databases and creds) the package and
reinstalling, with a similar result.

That's a rather strange error - a 767 bytes key does seem very long. ;) I
suspect something might be going on in dbconfig-common - the mailman postinst
doesn't seem to have anything specific to this, for example. Note that
dbconfig-common was upgraded to backports as a dependency, so this might be
where the trouble lies as well.

Unfortunately, I have about zero clue how to debug dbconfig, which has always
been a horrible black box to me, so I am sorry I cannot help further. 

-- System Information:
Debian Release: 9.7
  APT prefers stable
  APT policy: (500, 'stable'), (1, 'testing')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.9.0-8-amd64 (SMP w/2 CPU cores)
Locale: LANG=fr_CA.UTF-8, LC_CTYPE=fr_CA.UTF-8 (charmap=UTF-8), LANGUAGE=fr_CA.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages mailman3-web depends on:
ii  dbconfig-sqlite3           2.0.11~bpo9+1
ii  debconf [debconf-2.0]      1.5.61
ii  lsb-base                   9.20161125
ii  node-less                  1.6.3~dfsg-2
ii  python3                    3.5.3-1
ii  python3-django-hyperkitty  1.2.1-4~bpo9+1
ii  python3-django-postorius   1.2.2-4~bpo9+2
ii  python3-psycopg2           2.6.2-1
ii  python3-pymysql            0.7.10-1
ii  python3-whoosh             2.7.0-2
ii  ruby-sass                  3.5.3-1~bpo9+1
ii  ucf                        3.0036
ii  uwsgi                      2.0.14+20161117-3+deb9u2
ii  uwsgi-plugin-python3       2.0.14+20161117-3+deb9u2

Versions of packages mailman3-web recommends:
ii  libapache2-mod-proxy-uwsgi  2.0.14+20161117-3+deb9u2
ii  nginx                       1.10.3-1+deb9u2
ii  nginx-full [nginx]          1.10.3-1+deb9u2

Versions of packages mailman3-web suggests:
ii  mariadb-server-10.1 [virtual-mysql-server]  10.1.37-0+deb9u1

-- debconf information:
  mailman3-web/db/basepath: /var/lib/mailman3/web
  mailman3-web/django-site: anarc.at
  mailman3-web/dbconfig-reinstall: false
  mailman3-web/nginx-choice:
  mailman3-web/upgrade-error: abort
  mailman3-web/remove-error: abort
  mailman3-web/mysql/method: Unix socket
  mailman3-web/db/app-user: mailman3web@localhost
  mailman3-web/configure-webserver: none
  mailman3-web/dbconfig-upgrade: true
  mailman3-web/dbconfig-remove: true
  mailman3-web/internal/skip-preseed: false
  mailman3-web/passwords-do-not-match:
  mailman3-web/remote/port:
  mailman3-web/pgsql/authmethod-user: password
  mailman3-web/pgsql/manualconf:
  mailman3-web/superuser-name: admin
* mailman3-web/database-type: mysql
  mailman3-web/purge: false
  mailman3-web/missing-db-package-error: abort
  mailman3-web/internal/reconfiguring: false
  mailman3-web/pgsql/method: TCP/IP
  mailman3-web/install-error: abort
  mailman3-web/superuser-mail: root@localhost
  mailman3-web/emailname: localhost.local
  mailman3-web/restart-webserver: true
  mailman3-web/pgsql/changeconf: false
  mailman3-web/db/dbname: mailman3web
  mailman3-web/pgsql/authmethod-admin: ident
  mailman3-web/upgrade-backup: true
* mailman3-web/mysql/admin-user: debian-sys-maint
  mailman3-web/remote/newhost:
  mailman3-web/pgsql/no-empty-passwords:
  mailman3-web/pgsql/admin-user: postgres
* mailman3-web/dbconfig-install: true
  mailman3-web/remote/host: localhost

Message #10 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Antoine Beaupre <anarcat@debian.org>

To: Debian Bug Tracking System <921128@bugs.debian.org>

Subject: Re: mailman3-web fails to initialize mysql: Specified key was too long

Date: Fri, 01 Feb 2019 19:02:32 -0500

Package: mailman3-web
Followup-For: Bug #921128

I have tried to reproduce this in buster and at first I seem to recall
I did reproduce it, but now I somewhat managed to get through and have
it installed correctly.

Also note this might be a bug specific to MySQL: running the dbconfig
stuff with a sqlite3 backend doesn't trigger the same bug, even in
backports.

That said, the bug doesn't occur within dbconfig itself. It's the
django migration that raises the backtrace (which is logical
considering dbconfig is written in shell and not
Python). Specifically, this command reproduces the problem outside of
dpkg:

su --shell /bin/sh --command "python3 /usr/bin/django-admin migrate --no-input --verbosity 3 --pythonpath /usr/share/mailman3-web --settings settings" www-data

And, for what it's worth, I tried backporting the 0+20180916-4 release
of mailman-suite, to no effect: the same error still occurs.

A.

-- System Information:
Debian Release: buster/sid
  APT prefers testing
  APT policy: (500, 'testing'), (1, 'experimental'), (1, 'unstable')
Architecture: amd64 (x86_64)

Kernel: Linux 4.19.0-1-amd64 (SMP w/4 CPU cores)
Locale: LANG=fr_CA.UTF-8, LC_CTYPE=fr_CA.UTF-8 (charmap=UTF-8), LANGUAGE=fr_CA.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages mailman3-web depends on:
ii  dbconfig-sqlite3           2.0.11
ii  debconf [debconf-2.0]      1.5.70
ii  lsb-base                   10.2018112800
ii  node-less                  1.6.3~dfsg-3
ii  python3                    3.7.2-1
ii  python3-django-hyperkitty  1.2.1-4
ii  python3-django-postorius   1.2.2-4
ii  python3-psycopg2           2.7.7-1
ii  python3-pymysql            0.9.3-1
ii  python3-whoosh             2.7.4+git6-g9134ad92-1
ii  ruby-sass                  3.5.6-1
ii  ucf                        3.0038+nmu1
ii  uwsgi                      2.0.17.1-11
ii  uwsgi-plugin-python3       2.0.17.1-11

Versions of packages mailman3-web recommends:
ii  libapache2-mod-proxy-uwsgi  2.4.37-1

Versions of packages mailman3-web suggests:
ii  mariadb-server-10.3 [virtual-mysql-server]  1:10.3.12-2

-- debconf-show failed

Message #15 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Antoine Beaupré <anarcat@debian.org>

To: 921128@bugs.debian.org

Subject: Re: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Fri, 01 Feb 2019 21:05:38 -0500

I just read the README.Debian file and it says the mariadb version in
stretch might conflict with the mailman3-web version.

If that's really the case, might I suggest the backport be fixed to warn
explicitely about this somehow? maybe conflict with that mariadb
version?

A.
-- 
Software gets slower faster than hardware gets faster.
                         - Wirth's law

Message #20 received at 921128-submitter@bugs.debian.org (full text, mbox, reply):

From: Pierre-Elliott Bécue <>, @godard.debian.org

To: 921128-submitter@bugs.debian.org

Subject: Bug #921128 in mailman-suite marked as pending

Date: Sun, 10 Feb 2019 00:09:38 +0000

Control: tag -1 pending

Hello,

Bug #921128 in mailman-suite reported by you has been fixed in the
Git repository and is awaiting an upload. You can see the commit
message below and you can check the diff of the fix at:

https://salsa.debian.org/mailman-team/mailman-suite/commit/1dc0dcf43e763b4b78e808877d65a8dbb6119170

------------------------------------------------------------------------
d/mailman3-web.preinst: Add an explicit warning regarding the conflict between mailman3-web and MariaDB

Closes: #921128
------------------------------------------------------------------------

(this message was generated automatically)
-- 
Greetings

https://bugs.debian.org/921128

Message #27 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Pierre-Elliott Bécue <peb@debian.org>

To: Antoine Beaupré <anarcat@debian.org>, 921128@bugs.debian.org

Subject: Re: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Sun, 10 Feb 2019 01:11:35 +0100

[Message part 1 (text/plain, inline)]

Le vendredi 01 février 2019 à 21:05:38-0500, Antoine Beaupré a écrit :
> I just read the README.Debian file and it says the mariadb version in
> stretch might conflict with the mailman3-web version.
> 
> If that's really the case, might I suggest the backport be fixed to warn
> explicitely about this somehow? maybe conflict with that mariadb
> version?

Please review and comment
https://salsa.debian.org/mailman-team/mailman-suite/commit/1dc0dcf43e763b4b78e808877d65a8dbb6119170

I'll upload in a couple of days. :)

-- 
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528  F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.

[signature.asc (application/pgp-signature, inline)]

Message #32 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Antoine Beaupré <anarcat@debian.org>

To: Pierre-Elliott Bécue <peb@debian.org>, 921128@bugs.debian.org

Subject: Re: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Sun, 10 Feb 2019 18:10:55 -0500

On 2019-02-10 01:11:35, Pierre-Elliott Bécue wrote:
> Le vendredi 01 février 2019 à 21:05:38-0500, Antoine Beaupré a écrit :
>> I just read the README.Debian file and it says the mariadb version in
>> stretch might conflict with the mailman3-web version.
>> 
>> If that's really the case, might I suggest the backport be fixed to warn
>> explicitely about this somehow? maybe conflict with that mariadb
>> version?
>
> Please review and comment
> https://salsa.debian.org/mailman-team/mailman-suite/commit/1dc0dcf43e763b4b78e808877d65a8dbb6119170
>
> I'll upload in a couple of days. :)

That's a great start!

Couldn't we check if mariadb is actually installed or configured somehow
instead of just always prompting?

But if that's too much work, it's better than nothing for sure.

THanks! :)

A.

-- 
On ne résout pas un problème avec les modes de pensée qui l'ont
engendré.
                        - Albert Einstein

Message #37 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Pierre-Elliott Bécue <peb@debian.org>

To: Antoine Beaupré <anarcat@debian.org>,921128@bugs.debian.org

Subject: Re: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Mon, 11 Feb 2019 09:01:22 +0100

Le 11 février 2019 00:10:55 GMT+01:00, "Antoine Beaupré" <anarcat@debian.org> a écrit :
>On 2019-02-10 01:11:35, Pierre-Elliott Bécue wrote:
>> Le vendredi 01 février 2019 à 21:05:38-0500, Antoine Beaupré a
>écrit :
>>> I just read the README.Debian file and it says the mariadb version
>in
>>> stretch might conflict with the mailman3-web version.
>>> 
>>> If that's really the case, might I suggest the backport be fixed to
>warn
>>> explicitely about this somehow? maybe conflict with that mariadb
>>> version?
>>
>> Please review and comment
>>
>https://salsa.debian.org/mailman-team/mailman-suite/commit/1dc0dcf43e763b4b78e808877d65a8dbb6119170
>>
>> I'll upload in a couple of days. :)
>
>That's a great start!
>
>Couldn't we check if mariadb is actually installed or configured
>somehow
>instead of just always prompting?
>
>But if that's too much work, it's better than nothing for sure.
>
>THanks! :)
>
>A.

The risk is that if the user plans to use an external DB you won't warn him.

Cheers ! 
-- 
PEB from my phone.

Message #42 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Antoine Beaupré <anarcat@debian.org>

To: Pierre-Elliott Bécue <peb@debian.org>, 921128@bugs.debian.org

Subject: Re: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Mon, 11 Feb 2019 09:14:55 -0500

On 2019-02-11 09:01:22, Pierre-Elliott Bécue wrote:
> Le 11 février 2019 00:10:55 GMT+01:00, "Antoine Beaupré" <anarcat@debian.org> a écrit :
>>On 2019-02-10 01:11:35, Pierre-Elliott Bécue wrote:
>>> Le vendredi 01 février 2019 à 21:05:38-0500, Antoine Beaupré a
>>écrit :
>>>> I just read the README.Debian file and it says the mariadb version
>>in
>>>> stretch might conflict with the mailman3-web version.
>>>> 
>>>> If that's really the case, might I suggest the backport be fixed to
>>warn
>>>> explicitely about this somehow? maybe conflict with that mariadb
>>>> version?
>>>
>>> Please review and comment
>>>
>>https://salsa.debian.org/mailman-team/mailman-suite/commit/1dc0dcf43e763b4b78e808877d65a8dbb6119170
>>>
>>> I'll upload in a couple of days. :)
>>
>>That's a great start!
>>
>>Couldn't we check if mariadb is actually installed or configured
>>somehow
>>instead of just always prompting?
>>
>>But if that's too much work, it's better than nothing for sure.
>>
>>THanks! :)
>>
>>A.
>
> The risk is that if the user plans to use an external DB you won't warn him.

Right, the only way this could be properly implemented would be if
debconf would attempt to connect with the configured DB credentials to
check the remote server version (or, better, to test the actuall code
bugginess).

That's probably overkill though.

A.

-- 
The future is already here – it's just not very evenly distributed.
                       - William Gibson

Message #47 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Pierre-Elliott Bécue <peb@debian.org>

To: Antoine Beaupré <anarcat@debian.org>

Cc: 921128@bugs.debian.org

Subject: Re: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Mon, 11 Feb 2019 21:52:32 +0100

[Message part 1 (text/plain, inline)]

Le lundi 11 février 2019 à 09:14:55-0500, Antoine Beaupré a écrit :
> On 2019-02-11 09:01:22, Pierre-Elliott Bécue wrote:
> > Le 11 février 2019 00:10:55 GMT+01:00, "Antoine Beaupré" <anarcat@debian.org> a écrit :
> >>On 2019-02-10 01:11:35, Pierre-Elliott Bécue wrote:
> >>> Le vendredi 01 février 2019 à 21:05:38-0500, Antoine Beaupré a
> >>écrit :
> >>>> I just read the README.Debian file and it says the mariadb version
> >>in
> >>>> stretch might conflict with the mailman3-web version.
> >>>> 
> >>>> If that's really the case, might I suggest the backport be fixed to
> >>warn
> >>>> explicitely about this somehow? maybe conflict with that mariadb
> >>>> version?
> >>>
> >>> Please review and comment
> >>>
> >>https://salsa.debian.org/mailman-team/mailman-suite/commit/1dc0dcf43e763b4b78e808877d65a8dbb6119170
> >>>
> >>> I'll upload in a couple of days. :)
> >>
> >>That's a great start!
> >>
> >>Couldn't we check if mariadb is actually installed or configured
> >>somehow
> >>instead of just always prompting?
> >>
> >>But if that's too much work, it's better than nothing for sure.
> >>
> >>THanks! :)
> >>
> >>A.
> >
> > The risk is that if the user plans to use an external DB you won't warn him.
> 
> Right, the only way this could be properly implemented would be if
> debconf would attempt to connect with the configured DB credentials to
> check the remote server version (or, better, to test the actuall code
> bugginess).
> 
> That's probably overkill though.

ISTM that it is.

But if you wish to go to such extents I'd be glad to include a patch!
(no time to dev it though)

Chaars,

-- 
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528  F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.

[signature.asc (application/pgp-signature, inline)]

Message #52 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Antoine Beaupré <anarcat@debian.org>

To: Pierre-Elliott Bécue <peb@debian.org>

Cc: 921128@bugs.debian.org

Subject: Re: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Mon, 11 Feb 2019 16:27:21 -0500

On 2019-02-11 21:52:32, Pierre-Elliott Bécue wrote:
> Le lundi 11 février 2019 à 09:14:55-0500, Antoine Beaupré a écrit :
>> On 2019-02-11 09:01:22, Pierre-Elliott Bécue wrote:
>> > Le 11 février 2019 00:10:55 GMT+01:00, "Antoine Beaupré" <anarcat@debian.org> a écrit :
>> >>On 2019-02-10 01:11:35, Pierre-Elliott Bécue wrote:
>> >>> Le vendredi 01 février 2019 à 21:05:38-0500, Antoine Beaupré a
>> >>écrit :
>> >>>> I just read the README.Debian file and it says the mariadb version
>> >>in
>> >>>> stretch might conflict with the mailman3-web version.
>> >>>> 
>> >>>> If that's really the case, might I suggest the backport be fixed to
>> >>warn
>> >>>> explicitely about this somehow? maybe conflict with that mariadb
>> >>>> version?
>> >>>
>> >>> Please review and comment
>> >>>
>> >>https://salsa.debian.org/mailman-team/mailman-suite/commit/1dc0dcf43e763b4b78e808877d65a8dbb6119170
>> >>>
>> >>> I'll upload in a couple of days. :)
>> >>
>> >>That's a great start!
>> >>
>> >>Couldn't we check if mariadb is actually installed or configured
>> >>somehow
>> >>instead of just always prompting?
>> >>
>> >>But if that's too much work, it's better than nothing for sure.
>> >>
>> >>THanks! :)
>> >>
>> >>A.
>> >
>> > The risk is that if the user plans to use an external DB you won't warn him.
>> 
>> Right, the only way this could be properly implemented would be if
>> debconf would attempt to connect with the configured DB credentials to
>> check the remote server version (or, better, to test the actuall code
>> bugginess).
>> 
>> That's probably overkill though.
>
> ISTM that it is.
>
> But if you wish to go to such extents I'd be glad to include a patch!
> (no time to dev it though)

Alright, let good win in the face of non-existing perfect threats. ;)

A.

-- 
During the initial stage of the struggle, the oppressed, instead of
striving for liberation, tend themselves to become oppressors The very
structure of their thought has been conditioned by the contradictions of
the concrete, existential situation by which they were shaped. Their
ideal is to be men; but for them, to be men is to be oppressors. This is
their model of humanity.
                        - Paulo Freire, Pedagogy of the Oppressed

Message #57 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Pierre-Elliott Bécue <peb@debian.org>

To: Antoine Beaupré <anarcat@debian.org>

Cc: 921128@bugs.debian.org

Subject: Re: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Mon, 11 Feb 2019 22:53:51 +0100

[Message part 1 (text/plain, inline)]

Le lundi 11 février 2019 à 16:27:21-0500, Antoine Beaupré a écrit :
> On 2019-02-11 21:52:32, Pierre-Elliott Bécue wrote:
> > Le lundi 11 février 2019 à 09:14:55-0500, Antoine Beaupré a écrit :
> >> Right, the only way this could be properly implemented would be if
> >> debconf would attempt to connect with the configured DB credentials to
> >> check the remote server version (or, better, to test the actuall code
> >> bugginess).
> >> 
> >> That's probably overkill though.
> >
> > ISTM that it is.
> >
> > But if you wish to go to such extents I'd be glad to include a patch!
> > (no time to dev it though)
> 
> Alright, let good win in the face of non-existing perfect threats. ;)

I'll leave ourselves 24h of thinking and then I'll upload.

If you get attacked by insomnia, you know what to do. :P

-- 
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528  F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.

[signature.asc (application/pgp-signature, inline)]

Message #62 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Jonas Meurer <jonas@freesources.org>

To: Pierre-Elliott Bécue <peb@debian.org>, 921128@bugs.debian.org, Antoine Beaupré <anarcat@debian.org>

Subject: Re: [Pkg-mailman-hackers] Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Mon, 11 Feb 2019 23:40:51 +0100

[Message part 1 (text/plain, inline)]

Hi PEB, hi anarcat,

Pierre-Elliott Bécue:
> Le lundi 11 février 2019 à 16:27:21-0500, Antoine Beaupré a écrit :
>> On 2019-02-11 21:52:32, Pierre-Elliott Bécue wrote:
>>> Le lundi 11 février 2019 à 09:14:55-0500, Antoine Beaupré a écrit :
>>>> Right, the only way this could be properly implemented would be if
>>>> debconf would attempt to connect with the configured DB credentials to
>>>> check the remote server version (or, better, to test the actuall code
>>>> bugginess).
>>>>
>>>> That's probably overkill though.
>>>
>>> ISTM that it is.
>>>
>>> But if you wish to go to such extents I'd be glad to include a patch!
>>> (no time to dev it though)
>>
>> Alright, let good win in the face of non-existing perfect threats. ;)
> 
> I'll leave ourselves 24h of thinking and then I'll upload.
> 
> If you get attacked by insomnia, you know what to do. :P

Sorry for not responding earlier, I was to busy with other things.
Still, I should have reacted earlier ...

As you might have seen, mariadb-10.1 10.1.37-0+deb9u1 is in stable-new
queue[1]. I don't know the details why it didn't proceed to stable yet,
but there's hope that this issue will be sorted out in Stretch soon. The
10.1.37 contains both the upstream backport of
`innodb_default_row_format`[2] and the patch to the config[3] that sets
the default row format to `dynamic`, so the `max key length is 767
bytes` error is definitely fixed in this release.

Therefore I don't think it's still *needed* to upload the patch that PEB
poposed in this bugreport. I don't *object* to uploading it either. So
if you want to, go ahead, PEB. But we should track the mariadb 10.1
version in Stretch and remove the Debconf question again once mariadb
10.1.37 got accepted into Stretch.

Cheers
 jonas

[1]
https://tracker.debian.org/news/1004285/accepted-mariadb-101-10137-0deb9u1-source-into-stable-embargoed-stable/
[2] https://jira.mariadb.org/browse/MDEV-14904
[3]
https://salsa.debian.org/mariadb-team/mariadb-10.1/commit/89ae638d1d3b5a7157d086a9be2468cae764aae7

[signature.asc (application/pgp-signature, attachment)]

Message #67 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Pierre-Elliott Bécue <peb@debian.org>

To: Jonas Meurer <jonas@freesources.org>, 921128@bugs.debian.org

Cc: Antoine Beaupré <anarcat@debian.org>

Subject: Re: [Pkg-mailman-hackers] Bug#921128: Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Tue, 12 Feb 2019 10:37:25 +0100

Le lundi 11 février 2019 à 23:40:51+0100, Jonas Meurer a écrit :
> Hi PEB, hi anarcat,
> 
> Pierre-Elliott Bécue:
> > Le lundi 11 février 2019 à 16:27:21-0500, Antoine Beaupré a écrit :
> >> On 2019-02-11 21:52:32, Pierre-Elliott Bécue wrote:
> >>> Le lundi 11 février 2019 à 09:14:55-0500, Antoine Beaupré a écrit :
> >>>> Right, the only way this could be properly implemented would be if
> >>>> debconf would attempt to connect with the configured DB credentials to
> >>>> check the remote server version (or, better, to test the actuall code
> >>>> bugginess).
> >>>>
> >>>> That's probably overkill though.
> >>>
> >>> ISTM that it is.
> >>>
> >>> But if you wish to go to such extents I'd be glad to include a patch!
> >>> (no time to dev it though)
> >>
> >> Alright, let good win in the face of non-existing perfect threats. ;)
> > 
> > I'll leave ourselves 24h of thinking and then I'll upload.
> > 
> > If you get attacked by insomnia, you know what to do. :P
> 
> Sorry for not responding earlier, I was to busy with other things.
> Still, I should have reacted earlier ...
> 
> As you might have seen, mariadb-10.1 10.1.37-0+deb9u1 is in stable-new
> queue[1]. I don't know the details why it didn't proceed to stable yet,

Because any release in stable is done manually by the stable managers. I
think it won't get included before a minor release at least, and we
can't be sure as they're free to pick (or not) packages from the stable
proposed updates.

> but there's hope that this issue will be sorted out in Stretch soon. The
> 10.1.37 contains both the upstream backport of
> `innodb_default_row_format`[2] and the patch to the config[3] that sets
> the default row format to `dynamic`, so the `max key length is 767
> bytes` error is definitely fixed in this release.
> 
> Therefore I don't think it's still *needed* to upload the patch that PEB
> poposed in this bugreport. I don't *object* to uploading it either. So
> if you want to, go ahead, PEB. But we should track the mariadb 10.1
> version in Stretch and remove the Debconf question again once mariadb
> 10.1.37 got accepted into Stretch.

I'll let the warning for now and remove it as soon as we know the
package is included.

-- 
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528  F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.

Message #72 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Pierre-Elliott Bécue <peb@debian.org>

To: Jonas Meurer <jonas@freesources.org>

Cc: 921128@bugs.debian.org, Antoine Beaupré <anarcat@debian.org>

Subject: Re: [Pkg-mailman-hackers] Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Wed, 13 Feb 2019 21:51:12 +0100

[Message part 1 (text/plain, inline)]

Le lundi 11 février 2019 à 23:40:51+0100, Jonas Meurer a écrit :
> Hi PEB, hi anarcat,
> 
> Pierre-Elliott Bécue:
> > Le lundi 11 février 2019 à 16:27:21-0500, Antoine Beaupré a écrit :
> >> On 2019-02-11 21:52:32, Pierre-Elliott Bécue wrote:
> >>> Le lundi 11 février 2019 à 09:14:55-0500, Antoine Beaupré a écrit :
> >>>> Right, the only way this could be properly implemented would be if
> >>>> debconf would attempt to connect with the configured DB credentials to
> >>>> check the remote server version (or, better, to test the actuall code
> >>>> bugginess).
> >>>>
> >>>> That's probably overkill though.
> >>>
> >>> ISTM that it is.
> >>>
> >>> But if you wish to go to such extents I'd be glad to include a patch!
> >>> (no time to dev it though)
> >>
> >> Alright, let good win in the face of non-existing perfect threats. ;)
> > 
> > I'll leave ourselves 24h of thinking and then I'll upload.
> > 
> > If you get attacked by insomnia, you know what to do. :P
> 
> Sorry for not responding earlier, I was to busy with other things.
> Still, I should have reacted earlier ...
> 
> As you might have seen, mariadb-10.1 10.1.37-0+deb9u1 is in stable-new
> queue[1]. I don't know the details why it didn't proceed to stable yet,
> but there's hope that this issue will be sorted out in Stretch soon. The
> 10.1.37 contains both the upstream backport of
> `innodb_default_row_format`[2] and the patch to the config[3] that sets
> the default row format to `dynamic`, so the `max key length is 767
> bytes` error is definitely fixed in this release.
> 
> Therefore I don't think it's still *needed* to upload the patch that PEB
> poposed in this bugreport. I don't *object* to uploading it either. So
> if you want to, go ahead, PEB. But we should track the mariadb 10.1
> version in Stretch and remove the Debconf question again once mariadb
> 10.1.37 got accepted into Stretch.
> 
> Cheers

This mariadb release seems to be already in security for stretch.

I changed a little the warning but I think we should keep it (at least
until the point release of stretch is really done):
https://salsa.debian.org/mailman-team/mailman-suite/commits/stretch-backports

Tell me what you think about it.

-- 
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528  F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.

[signature.asc (application/pgp-signature, inline)]

Message #77 received at 921128@bugs.debian.org (full text, mbox, reply):

From: Jonas Meurer <jonas@freesources.org>

To: Pierre-Elliott Bécue <peb@debian.org>

Cc: 921128@bugs.debian.org, Antoine Beaupré <anarcat@debian.org>

Subject: Re: [Pkg-mailman-hackers] Bug#921128: Info received (mailman3-web fails to initialize mysql: Specified key was too long)

Date: Fri, 15 Feb 2019 00:57:29 +0100

[Message part 1 (text/plain, inline)]

Pierre-Elliott Bécue:
> Le lundi 11 février 2019 à 23:40:51+0100, Jonas Meurer a écrit :
>> Hi PEB, hi anarcat,
>>
>> Pierre-Elliott Bécue:
>>> Le lundi 11 février 2019 à 16:27:21-0500, Antoine Beaupré a écrit :
>>>> On 2019-02-11 21:52:32, Pierre-Elliott Bécue wrote:
>>>>> Le lundi 11 février 2019 à 09:14:55-0500, Antoine Beaupré a écrit :
>>>>>> Right, the only way this could be properly implemented would be if
>>>>>> debconf would attempt to connect with the configured DB credentials to
>>>>>> check the remote server version (or, better, to test the actuall code
>>>>>> bugginess).
>>>>>>
>>>>>> That's probably overkill though.
>>>>>
>>>>> ISTM that it is.
>>>>>
>>>>> But if you wish to go to such extents I'd be glad to include a patch!
>>>>> (no time to dev it though)
>>>>
>>>> Alright, let good win in the face of non-existing perfect threats. ;)
>>>
>>> I'll leave ourselves 24h of thinking and then I'll upload.
>>>
>>> If you get attacked by insomnia, you know what to do. :P
>>
>> Sorry for not responding earlier, I was to busy with other things.
>> Still, I should have reacted earlier ...
>>
>> As you might have seen, mariadb-10.1 10.1.37-0+deb9u1 is in stable-new
>> queue[1]. I don't know the details why it didn't proceed to stable yet,
>> but there's hope that this issue will be sorted out in Stretch soon. The
>> 10.1.37 contains both the upstream backport of
>> `innodb_default_row_format`[2] and the patch to the config[3] that sets
>> the default row format to `dynamic`, so the `max key length is 767
>> bytes` error is definitely fixed in this release.
>>
>> Therefore I don't think it's still *needed* to upload the patch that PEB
>> poposed in this bugreport. I don't *object* to uploading it either. So
>> if you want to, go ahead, PEB. But we should track the mariadb 10.1
>> version in Stretch and remove the Debconf question again once mariadb
>> 10.1.37 got accepted into Stretch.
>>
>> Cheers
> 
> This mariadb release seems to be already in security for stretch.
> 
> I changed a little the warning but I think we should keep it (at least
> until the point release of stretch is really done):
> https://salsa.debian.org/mailman-team/mailman-suite/commits/stretch-backports
> 
> Tell me what you think about it.

Thanks for taking care, PEB! As written, I don't have strong feelings on
that, given that a compatible MariaDB should propagate to Stretch soon
(hopefully). Feel free to upload as is, I reviewed the commits and think
they're fine.

I agree with you though that we should remove the debconf notification
again once the compatible MariaDB packages went into Stretch.

Cheers
 jonas

> 

[signature.asc (application/pgp-signature, attachment)]

Message #82 received at 921128-done@bugs.debian.org (full text, mbox, reply):

From: Jonas Meurer <jonas@freesources.org>

To: Antoine Beaupre <anarcat@debian.org>, 921128-done@bugs.debian.org

Subject: Re: [Pkg-mailman-hackers] Bug#921128: mailman3-web fails to initialize mysql: Specified key was too long

Date: Sat, 23 Jan 2021 21:58:18 +0100

[Message part 1 (text/plain, inline)]

Hello,

closing this bugreport as stretch-backports has been discontinued since 
quite some time.

Kind regards
 jonas

Am 02.02.19 um 01:02 schrieb Antoine Beaupre:
> Package: mailman3-web
> Followup-For: Bug #921128
> 
> I have tried to reproduce this in buster and at first I seem to recall
> I did reproduce it, but now I somewhat managed to get through and have
> it installed correctly.
> 
> Also note this might be a bug specific to MySQL: running the dbconfig
> stuff with a sqlite3 backend doesn't trigger the same bug, even in
> backports.
> 
> That said, the bug doesn't occur within dbconfig itself. It's the
> django migration that raises the backtrace (which is logical
> considering dbconfig is written in shell and not
> Python). Specifically, this command reproduces the problem outside of
> dpkg:
> 
> su --shell /bin/sh --command "python3 /usr/bin/django-admin migrate --no-input --verbosity 3 --pythonpath /usr/share/mailman3-web --settings settings" www-data
> 
> And, for what it's worth, I tried backporting the 0+20180916-4 release
> of mailman-suite, to no effect: the same error still occurs.
> 
> A.
> 
> -- System Information:
> Debian Release: buster/sid
>    APT prefers testing
>    APT policy: (500, 'testing'), (1, 'experimental'), (1, 'unstable')
> Architecture: amd64 (x86_64)
> 
> Kernel: Linux 4.19.0-1-amd64 (SMP w/4 CPU cores)
> Locale: LANG=fr_CA.UTF-8, LC_CTYPE=fr_CA.UTF-8 (charmap=UTF-8), LANGUAGE=fr_CA.UTF-8 (charmap=UTF-8)
> Shell: /bin/sh linked to /usr/bin/dash
> Init: systemd (via /run/systemd/system)
> LSM: AppArmor: enabled
> 
> Versions of packages mailman3-web depends on:
> ii  dbconfig-sqlite3           2.0.11
> ii  debconf [debconf-2.0]      1.5.70
> ii  lsb-base                   10.2018112800
> ii  node-less                  1.6.3~dfsg-3
> ii  python3                    3.7.2-1
> ii  python3-django-hyperkitty  1.2.1-4
> ii  python3-django-postorius   1.2.2-4
> ii  python3-psycopg2           2.7.7-1
> ii  python3-pymysql            0.9.3-1
> ii  python3-whoosh             2.7.4+git6-g9134ad92-1
> ii  ruby-sass                  3.5.6-1
> ii  ucf                        3.0038+nmu1
> ii  uwsgi                      2.0.17.1-11
> ii  uwsgi-plugin-python3       2.0.17.1-11
> 
> Versions of packages mailman3-web recommends:
> ii  libapache2-mod-proxy-uwsgi  2.4.37-1
> 
> Versions of packages mailman3-web suggests:
> ii  mariadb-server-10.3 [virtual-mysql-server]  1:10.3.12-2
> 
> -- debconf-show failed
> 
> _______________________________________________
> Pkg-mailman-hackers mailing list
> Pkg-mailman-hackers@alioth-lists.debian.net
> https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/pkg-mailman-hackers
> 

[OpenPGP_signature (application/pgp-signature, attachment)]

Send a report that this bug log contains spam.