Op 08-11-2022 om 22:54 schreef Daniel Black:

On Wed, Nov 9, 2022 at 12:01 AM Jogchum Reitsma jogchum.reitsma@gmail.com wrote:

...

Op 08-11-2022 om 13:04 schreef Marko Mäkelä:

Hi Jogchum,

On Tue, Nov 8, 2022 at 1:21 PM Jogchum Reitsma jogchum.reitsma@gmail.com wrote: [...]

2022-11-08 12:06:17 0 [Warning] 'innodb-file-format' was removed. It does nothing now and exists only for compatibility with old my.cnf files. (repeated several times) 2022-11-08 12:06:17 0 [ERROR] /usr/local/mariadb-10.6.10-linux-systemd-x86_64/bin/mariadbd: unknown option '--skip-locking' 2022-11-08 12:06:17 0 [ERROR] Aborting

Was that the end of the output? No shutdown message from InnoDB, like this?

2022-11-08 14:03:03 0 [Note] InnoDB: Shutdown completed; log sequence number 2150107; transaction id 3085

If there was no output like that, you should remove any line that says "skip_locking" or "skip-locking" from the configuration file and retry.

Marko

Yes, that was indeed the end of the output.

The skip-locking statement was not in /etc/my.cnf, but inmy.cnf in my home dir. Commented it out and now this complaint disappeared.

But now I get

####################################################################

2022-11-08 13:37:32 0 [Note] InnoDB: Compressed tables use zlib 1.2.12 2022-11-08 13:37:32 0 [Note] InnoDB: Number of pools: 1 2022-11-08 13:37:32 0 [Note] InnoDB: Using crc32 + pclmulqdq instructions 2022-11-08 13:37:32 0 [Note] InnoDB: Using Linux native AIO 2022-11-08 13:37:32 0 [Note] InnoDB: Initializing buffer pool, total size = 134217728, chunk size = 134217728 2022-11-08 13:37:32 0 [Note] InnoDB: Completed initialization of buffer pool 2022-11-08 13:37:32 0 [Note] InnoDB: 128 rollback segments are active. 2022-11-08 13:37:32 0 [Note] InnoDB: Creating shared tablespace for temporary tables 2022-11-08 13:37:32 0 [Note] InnoDB: Setting file './ibtmp1' size to 12 MB. Physically writing the file full; Please wait ... 2022-11-08 13:37:32 0 [Note] InnoDB: File './ibtmp1' size is now 12 MB. 2022-11-08 13:37:32 0 [Note] InnoDB: 10.6.10 started; log sequence number 578191791; transaction id 2935782 2022-11-08 13:37:32 0 [Note] InnoDB: Loading buffer pool(s) from /home/jogchum/mysql/ib_buffer_pool 2022-11-08 13:37:32 0 [Note] Plugin 'FEEDBACK' is disabled. 2022-11-08 13:37:32 0 [Warning] 'innodb-file-format' was removed. It does nothing now and exists only for compatibility with old my.cnf files. 2022-11-08 13:37:33 0 [Note] Server socket created on IP: '0.0.0.0'. 2022-11-08 13:37:33 0 [ERROR] Can't start server : Bind on unix socket: No such file or directory

The directory /var/run/mysql doesn't exist

Not mentioned yet, the rights on /var/run/mysql are 755, so no problem there too. Only thing remarkable to me is that it has one file, called "protecteddir.", owned by root:root. No idea if this is of any significance to the problem

...

2022-11-08 13:37:33 0 [ERROR] Do you already have another server running on socket: /var/run/mysql/mysql.sock ?

This guess of an error should be fixed.

Am 10.11.22 um 12:06 schrieb Jogchum Reitsma:

Op 08-11-2022 om 22:54 schreef Daniel Black:

...

On Wed, Nov 9, 2022 at 12:01 AM Jogchum Reitsma jogchum.reitsma@gmail.com wrote:

...

<...> The directory /var/run/mysql doesn't exist

2022-11-08 13:37:33 0 [ERROR] Do you already have another server running on socket: /var/run/mysql/mysql.sock ?

This guess of an error should be fixed.

Well, in the meantime I have a running database, thanks to all the help I get here.

But I'm not quite out of trouble yet.

The last problem with starting /usr/local/mariadb-10.6.10-linux-systemd-x86_64/bin/mariadbd was solved by chmod 777 /var/run/mysql; this was needed because I start mariadb as a plain user for the moment. chmod on /var/log/msql gave me logging > Once again I'm at a loss here - what causes mysqld not to start?

Any further help dearly appreciated....

either "in the meantime I have a running database" or "what causes mysqld not to start" are possible and when it comes to "chmod on /var/log/msql gave me logging" show them

for the sake of god open "/usr/lib/systemd/system/mariadb.service" and remove or comment out the "ExecStartPre" until your problems are solved

"ExecStartPre=/usr/libexec/mysql/mysql-systemd-helper upgrade (code=exited, status=1/FAILURE)" is some distribution sepcific nonsense

in the real world it's enoug to ExecStart mysqld as user:ggroup mysql and skip all that nonsense including mysql_safe

PLEASE don't use "reply-all" on mailing-lists, you break "reply-list" for capable users!

Am 10.11.22 um 20:36 schrieb Jogchum Reitsma:

Op 10-11-2022 om 13:39 schreef Reindl Harald:

...

for the sake of god open "/usr/lib/systemd/system/mariadb.service" and remove or comment out the "ExecStartPre" until your problems are solved

"ExecStartPre=/usr/libexec/mysql/mysql-systemd-helper upgrade (code=exited, status=1/FAILURE)" is some distribution sepcific nonsense

I have no idea what these "install" and "upgrade " ExecStartPre do, but I commented them  out, did a systemctl daemon-reload, and issued systemctl start mariadb.service again.

Again with no luck:

systemctl status mariadb.service ×mariadb.service - MariaDB database server     Loaded: loaded (/usr/lib/systemd/system/mariadb.service; disabled; preset: disabled)    Drop-In: /etc/systemd/system/mariadb.service.d             └─override.conf     Active: failed(Result: exit-code) since Thu 2022-11-10 20:22:35 CET; 17s ago       Docs: man:mysqld(8) https://mariadb.com/kb/en/library/systemd/    Process: 35754 ExecStart=/usr/libexec/mysql/mysql-systemd-helper start (code=exited, status=1/FAILURE)   Main PID: 35754 (code=exited, status=1/FAILURE)     Status: "MariaDB server is down"

"ExecStart=/usr/libexec/mysql/mysql-systemd-helper" is still some distribution crap

throw away that crap by create a unit in "/etc/systemd/system/mariadb.service" which will completly override the distro stuff

* there is no need for helpers * there is no need for mysqld_safe and other nonsense * systemd can babysit services for years * that below is my mysqld/mriadb unit for years

"Type=notify" should work unless some moron compiled the binary, "Type=simple" would need helper scripts if services are ordered after the database

"After=network-up.service" should be replaced by whatever your distribution is using to start the network, as for most of the critical stuff i simply refuse to use the distribution nonsense and replaced newtork/iptables/ipset by my own services

"AmbientCapabilities=CAP_IPC_LOCK CAP_SYS_NICE" as well as "User=mysql" and "Group=mysql" are they key that the service can be started from the first second as the intended user (port 3306 don't need super user privileges) and all that crap around with all it's indirection makes debugging impossible and introduces security holes (there was a CVE in context mysqld_safe a few years ago)

minimize the crap used to start services to what is *really* needed

-----------

[Unit] Description=MariaDB Database After=network-up.service Before=crond.service ConditionPathExists=/etc/my.cnf

[Service] Type=notify KillMode=process KillSignal=SIGTERM SendSIGKILL=no

User=mysql Group=mysql

ExecStart=/usr/libexec/mysqld --defaults-file=/etc/my.cnf --pid-file=/dev/null Environment="LANG=C.UTF-8" Restart=always RestartSec=1 TimeoutSec=300 LimitNOFILE=infinity LimitMEMLOCK=infinity OOMScoreAdjust=-1000 TasksMax=2048

AmbientCapabilities=CAP_IPC_LOCK CAP_SYS_NICE CapabilityBoundingSet=CAP_IPC_LOCK CAP_SYS_NICE LockPersonality=yes MemoryDenyWriteExecute=yes NoNewPrivileges=yes PrivateDevices=yes PrivateTmp=yes ProtectClock=yes ProtectControlGroups=yes ProtectHome=yes ProtectHostname=yes ProtectKernelLogs=yes ProtectKernelModules=yes ProtectKernelTunables=yes RestrictAddressFamilies=AF_UNIX AF_LOCAL AF_INET AF_INET6 RestrictNamespaces=yes RestrictSUIDSGID=yes SystemCallArchitectures=native UMask=077

[Install] WantedBy=multi-user.target

Op 16-11-2022 om 01:49 schreef Daniel Black:

Read Marko's email again.

10.8+ won't start without a ib_logfile0. This is what not starting looks like.

Start it with at 10.7 version which will recreate ib_logfile0, and if you're lucky, you maybe won't have corrupted data because of this. Then try 10.8+.

I just corrected https://wiki.archlinux.org/title/MariaDB.

If you see other sites saying delete ib_logfile0, please share your experience with them

_______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp

OK thanks for your feedback.

The coming two weeks I'll be very busy with other projects, so it will take a time before I can get into this again.

AFAICS from  running the 10.6 version, the data is still good, and I've made a copy of the data-dir (besides of course normal back-ups).

It seems to me this version stuff is too sensitive: why won't 10.8  open a database which is running fine in 10.6? Should be some backwards compatibility, shouldn't it?

Any case, I get a lot of help here, so thanks for that!

regards, Jogchum

Op 06-12-2022 om 16:43 schreef Jogchum Reitsma:

Op 16-11-2022 om 01:49 schreef Daniel Black:

...

...

<...>

Read Marko's email again.

10.8+ won't start without a ib_logfile0. This is what not starting looks like.

Start it with at 10.7 version which will recreate ib_logfile0, and if you're lucky, you maybe won't have corrupted data because of this. Then try 10.8+.

I just correctedhttps://wiki.archlinux.org/title/MariaDB.

If you see other sites saying delete ib_logfile0, please share your experience with them

_______________________________________________ Mailing list:https://launchpad.net/~maria-discuss Post to :maria-discuss@lists.launchpad.net Unsubscribe :https://launchpad.net/~maria-discuss More help :https://help.launchpad.net/ListHelp

Hi,

So I downloaded and installed 10.7.7.

Bu now I get

/usr/local/mariadb-10.7.7-linux-systemd-x86_64/bin/mysql ERROR 2002 (HY000): Can't connect to local server through socket '/var/run/mysql/mysql.sock'

For now I have chmod /var/run/mysql to 777, so is writable by me when I try to start 10.7.7. But indeed the directory is empty.

In both /etc/my.cnf as ~/.my.cnf the socket is set to /var/run/mysql/mysql.sock

Hope you have some ideas to solve this one...

regards, Jogchum

Sorry, my bad: of course I should have started mysqld, not mysql.....

Starting 10.7.7 works, and indeed recreates ib_logfile0. All my databases are there, no corruption AFAICS.

Shut it down with kill -15, and the logfile reports a normal shutdown:

#####################################

cat mysqld.log 2022-12-07  9:29:20 0 [Note] InnoDB: Compressed tables use zlib 1.2.12 2022-12-07  9:29:20 0 [Note] InnoDB: Number of transaction pools: 1 2022-12-07  9:29:20 0 [Note] InnoDB: Using crc32 + pclmulqdq instructions 2022-12-07  9:29:20 0 [Note] InnoDB: Using Linux native AIO 2022-12-07  9:29:20 0 [Note] InnoDB: Initializing buffer pool, total size = 134217728, chunk size = 134217728 2022-12-07  9:29:20 0 [Note] InnoDB: Completed initialization of buffer pool 2022-12-07  9:29:21 0 [Note] InnoDB: 128 rollback segments are active. 2022-12-07  9:29:21 0 [Note] InnoDB: Creating shared tablespace for temporary tables 2022-12-07  9:29:21 0 [Note] InnoDB: Setting file './ibtmp1' size to 12 MB. Physically writing the file full; Please wait ... 2022-12-07  9:29:21 0 [Note] InnoDB: File './ibtmp1' size is now 12 MB. 2022-12-07  9:29:21 0 [Note] InnoDB: 10.7.7 started; log sequence number 578191915; transaction id 2935782 2022-12-07  9:29:21 0 [Note] InnoDB: Loading buffer pool(s) from /var/lib/mysql/ib_buffer_pool 2022-12-07  9:29:21 0 [Note] Plugin 'FEEDBACK' is disabled. 2022-12-07  9:29:21 0 [Warning] 'innodb-file-format' was removed. It does nothing now and exists only for compatibility with old my.cnf files. 2022-12-07  9:29:22 0 [Note] Server socket created on IP: '0.0.0.0'. 2022-12-07  9:29:22 0 [Note] /usr/local/mariadb-10.7.7-linux-systemd-x86_64/bin/mysqld: ready for connections. Version: '10.7.7-MariaDB-log'  socket: '/var/run/mysql/mysql.sock'  port: 3306  MariaDB Server 2022-12-07  9:29:23 0 [Note] InnoDB: Buffer pool(s) load completed at 221207  9:29:23 2022-12-07  9:39:23 0 [Note] /usr/local/mariadb-10.7.7-linux-systemd-x86_64/bin/mysqld (initiated by: unknown): Normal shutdown 2022-12-07  9:39:23 0 [Note] InnoDB: FTS optimize thread exiting. 2022-12-07  9:39:24 0 [Note] InnoDB: Starting shutdown... 2022-12-07  9:39:24 0 [Note] InnoDB: Dumping buffer pool(s) to /var/lib/mysql/ib_buffer_pool 2022-12-07  9:39:24 0 [Note] InnoDB: Restricted to 2028 pages due to innodb_buf_pool_dump_pct=25 2022-12-07  9:39:24 0 [Note] InnoDB: Buffer pool(s) dump completed at 221207  9:39:24 2022-12-07  9:39:24 0 [Note] InnoDB: Removed temporary tablespace data file: "./ibtmp1" 2022-12-07  9:39:24 0 [Note] InnoDB: Shutdown completed; log sequence number 578191927; transaction id 2935784 2022-12-07  9:39:24 0 [Note] /usr/local/mariadb-10.7.7-linux-systemd-x86_64/bin/mysqld: Shutdown complete

#######################################

So far so good.

But after that, still no luck in starting the server the regular way, i.e. with systemctl start mariadb.

######################################

systemctl start mariadb

Job for mariadb.service failed because the control process exited with error code. See "systemctl status mariadb.service" and "journalctl -xeu mariadb.service" for details

###########################################

systemctl status mariadb.service ×mariadb.service - MariaDB database server     Loaded: loaded (/usr/lib/systemd/system/mariadb.service; disabled; preset: disabled)    Drop-In: /etc/systemd/system/mariadb.service.d             └─override.conf     Active: failed(Result: exit-code) since Wed 2022-12-07 09:52:46 CET; 1min 51s ago       Docs: man:mysqld(8) https://mariadb.com/kb/en/library/systemd/    Process: 93046 ExecStartPre=/usr/libexec/mysql/mysql-systemd-helper install (code=exited, status=0/SUCCESS)    Process: 93059 ExecStartPre=/usr/libexec/mysql/mysql-systemd-helper upgrade (code=exited, status=1/FAILURE)        CPU: 438ms

dec 07 09:51:45 linux-mkay mysql-systemd-helper[93059]: Checking MySQL configuration for obsolete options... dec 07 09:51:45 linux-mkay mysql-systemd-helper[93059]: Trying to run upgrade of MySQL databases... dec 07 09:51:45 linux-mkay mysql-systemd-helper[93059]: Running protected MySQL... dec 07 09:51:45 linux-mkay mysql-systemd-helper[93059]: Waiting for MySQL to start dec 07 09:51:45 linux-mkay mysql-systemd-helper[93082]: 2022-12-07  9:51:45 0 [Note] /usr/sbin/mysqld (server 10.10.2-MariaDB) starting as process 93082 ... dec 07 09:52:46 linux-mkay mysql-systemd-helper[93059]: MySQL is still dead dec 07 09:52:46 linux-mkay mysql-systemd-helper[93059]: MySQL didn't start, can't continue dec 07 09:52:46 linux-mkay systemd[1]: mariadb.service: Control process exited, code=exited, status=1/FAILURE dec 07 09:52:46 linux-mkay systemd[1]: mariadb.service: Failed with result 'exit-code'. dec 07 09:52:46 linux-mkay systemd[1]: Failed to start MariaDB database server.

#################################################

journalctl -xeu mariadb.service

dec 07 09:51:45 linux-mkay systemd[1]: Starting MariaDB database server... ░░ Subject: A start job for unit mariadb.service has begun execution ░░Defined-By: systemd ░░Support: https://lists.freedesktop.org/mailman/listinfo/systemd-devel ░░ ░░A start job for unit mariadb.service has begun execution. ░░ ░░The job identifier is 611401. dec 07 09:51:45 linux-mkay mysql-systemd-helper[93059]: Checking MySQL configuration for obsolete options... dec 07 09:51:45 linux-mkay mysql-systemd-helper[93059]: Trying to run upgrade of MySQL databases... dec 07 09:51:45 linux-mkay mysql-systemd-helper[93059]: Running protected MySQL... dec 07 09:51:45 linux-mkay mysql-systemd-helper[93059]: Waiting for MySQL to start dec 07 09:51:45 linux-mkay mysql-systemd-helper[93082]: 2022-12-07  9:51:45 0 [Note] /usr/sbin/mysqld (server 10.10.2-MariaDB) starting as process 93082 ... dec 07 09:52:46 linux-mkay mysql-systemd-helper[93059]: MySQL is still dead dec 07 09:52:46 linux-mkay mysql-systemd-helper[93059]: MySQL didn't start, can't continue dec 07 09:52:46 linux-mkay systemd[1]: mariadb.service: Control process exited, code=exited, status=1/FAILURE ░░ Subject: Unit process exited ░░Defined-By: systemd ░░Support: https://lists.freedesktop.org/mailman/listinfo/systemd-devel ░░ ░░An ExecStartPre= process belonging to unit mariadb.service has exited. ░░ ░░The process' exit code is 'exited' and its exit status is 1. dec 07 09:52:46 linux-mkay systemd[1]: mariadb.service: Failed with result 'exit-code'. ░░ Subject: Unit failed ░░Defined-By: systemd ░░Support: https://lists.freedesktop.org/mailman/listinfo/systemd-devel ░░ ░░The unit mariadb.service has entered the 'failed' state with result 'exit-code'. dec 07 09:52:46 linux-mkay systemd[1]: Failed to start MariaDB database server. ░░ Subject: A start job for unit mariadb.service has failed ░░Defined-By: systemd ░░Support: https://lists.freedesktop.org/mailman/listinfo/systemd-devel ░░ ░░A start job for unit mariadb.service has finished with a failure. ░░ ░░The job identifier is 611401 and the job result is failed.

#########################################################################

The logfile is still the one created by 10.7.7, unaltered.

I don't see a clue to WHY MySQL didn't start in these reports.

Anyone any ideas here?

Of course, starting mysqld by hand, at least in the 10.7.7 version, gives me access to may data, but I would prefer starting the newest version in the standard systemd way...

Thanks in advance, again

regards, Jogchum

_______________________________________________ Mailing list:https://launchpad.net/~maria-discuss Post to :maria-discuss@lists.launchpad.net Unsubscribe :https://launchpad.net/~maria-discuss More help :https://help.launchpad.net/ListHelp

Op 07-12-2022 om 12:15 schreef Reindl Harald:

Am 07.12.22 um 10:27 schrieb Jogchum Reitsma:

...

Sorry, my bad: of course I should have started mysqld, not mysql.....

besides that you should notice when you try to start a non-existing service what is really bad: you randomly mix "mysqld", "mysql" and "mariadb" as service name in your posts

in the worst case they are really existing but doing different things

...

Starting 10.7.7 works, and indeed recreates ib_logfile0. All my databases are there, no corruption AFAICS.

Shut it down with kill -15, and the logfile reports a normal shutdown

you shouldn't use "kill" at all fro shutdown when you start something as systemd-unit

_______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to     : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help   : https://help.launchpad.net/ListHelp

Hi Reindl,

I  started the he 10.7.7 version not through systemd, but by issuing /usr/local/mariadb-10.7.7-linux-systemd-x86_64/bin/mysqld in the terminal, as advised earlier in this conversation. This way I managed to get a working db-server, and access to my data.

I stopped that by SIGQUIT, and this gives a correct shutdown.

After that, I tried to start the version that comes with the OS (OpenSuse Thumbleweed) trough systemctl start mariadb, without success.

Hope this clarifies the steps taken.

regards, Jogchum

Op 07-12-2022 om 23:14 schreef Reindl Harald:

Am 07.12.22 um 16:53 schrieb Jogchum Reitsma:

...

Op 07-12-2022 om 12:15 schreef Reindl Harald:

...

Am 07.12.22 um 10:27 schrieb Jogchum Reitsma:

...

Sorry, my bad: of course I should have started mysqld, not mysql.....

besides that you should notice when you try to start a non-existing service what is really bad: you randomly mix "mysqld", "mysql" and "mariadb" as service name in your posts

in the worst case they are really existing but doing different things

...

Starting 10.7.7 works, and indeed recreates ib_logfile0. All my databases are there, no corruption AFAICS.

Shut it down with kill -15, and the logfile reports a normal shutdown

you shouldn't use "kill" at all fro shutdown when you start something as systemd-unit

I  started the he 10.7.7 version not through systemd, but by issuing /usr/local/mariadb-10.7.7-linux-systemd-x86_64/bin/mysqld in the terminal, as advised earlier in this conversation.

as which user?

if it's a different one maybe permissions are a problem (again)

Well, as a normal user. But as I stated, there is no problem with *this* command: this works, and I get a functioning db-system.

"Sorry, my bad: of course I should have started mysqld, not mysql" should have been obvious - i mean a sevrer is a background process versus an interactive sql shell

That is what I meant to say, I inadvertently forgot the 'd' after 'mysql'.

...

This way I managed to get a working db-server, and access to my data.

I stopped that by SIGQUIT, and this gives a correct shutdown.

15 is SIGTERM, lucky you didn't find the number for SIGQUIT

I always use the numbers, not the variant in words.

...

After that, I tried to start the version that comes with the OS (OpenSuse Thumbleweed) trough systemctl start mariadb, without success.

so the problem is still not really solved?

if that's the case for the sake of god dump your data as soon as you get the server running,

As I said, I can get the server running, only not through systemd

remove all nonsense and mysql-data from the system, start from scratch and import the dumps as it's normal for postgresql

my mysql-data dirs date back to 2002 and where used on MacOS, Linux and Widows over the past years, so strange when mysql/mariadb acts that strange, but so be it

...

Hope this clarifies the steps taken

that whole thread is a mess....

_______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to     : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help   : https://help.launchpad.net/ListHelp

Hi Reindl,

Op 08-12-2022 om 08:31 schreef Reindl Harald:

FIRST: can you stop using "rewrite-all" on mailing-lists? when the useless offlist-copy is faster the dupes-filter on our server is supressing the list-copy and you break my reply-list button in thunderbird

OK, I'm sorry for that. But on some other mailing lists people ask explicitly to use reply all, so it's not always easy to do it the right way. But as you can see this goes only to the list.

I agree with you that this thread should com to an end, so I followed your advice, and dumped all my own tables, one by one, in <tablename>.sql files. Checked them, are OK.

<import the dumps as it's normal for postgresql>

I think you mean mysql, not postgresql...

I could dump table mysql as user root. The tables information_schema and performance_schema I could not dump. I suppose that is no problem?

If indeed that is no problem, I think the following steps are necessary:

- give ctrl-\ to the command running the 10.7.7 dbserver to gracefully stop it - umount the --bind mount to my database files - rename the directory where my database files are stored, - create a new, empty datadir in my home directory   ( I know that is not standard; we discussed it in this thread before.but I have several reasons to want it there.   After all, it has worked fine this way for some 20 years   Of course I always can mount --bind it to a more standard place, if  that is necessary or preferable) - delete the file structures created by untarring /usr/local/mariadb-10.6.10-linux-systemd-x86_64.tar.gz and /usr/local/mariadb-10.7.7-linux-systemd-x86_64.tar.gz - uninstall the mariadb package from the OS - remove /run/mysql (holds the lock file) - move /var/log/mysql to a place in my home system

Have I forgotten something?

After that, re-install the package from my OS. I would think the install process creates an empty database,  which I can populate with the dumps just created.

Thanks for all your replies, and sorry for the misunderstandings that made it so hard to help me. Of course that's never been my intention, I hope you can understand that.

that way this thread could come to an end

Hopefully it does...

regards Jogchum

_______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to     : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help   : https://help.launchpad.net/ListHelp

On Thu, Dec 8, 2022 at 5:41 PM Reindl Harald h.reindl@thelounge.net wrote:

...

<import the dumps as it's normal for postgresql> I think you mean mysql, not postgresql...

no, i mean postgresql, that piece of crap breaking for years after dist-upgrades because you need to do dump+restore at every version change - real fun because you need to remember about the dump *before*

The other side of the coin is a questionable InnoDB "optimization" that was finally fixed in MySQL 5.1.48 (in 2010) due to an external bug report. Heikki Tuuri firmly believed that zeroing out pages when they are initialized burns too many CPU cycles, so it is better to just write whatever garbage to any unused bytes in data pages. He actually forbade me to fix it back in 2004 or 2005, and forbade me to spend any time to prove what kind of bad things could happen. (Back then, even the FIL_PAGE_TYPE was written uninitialized to anything other than B-tree index pages. So, if the page type field said FIL_PAGE_INDEX, you could not say if it actually was an index page.)

I have tried to keep the implications in mind, but there has been at least one serious bug regarding this: https://jira.mariadb.org/browse/MDEV-27800

If a database with InnoDB had been originally initialized before MySQL 5.1.48 and you kept upgrading the binary files, you could have a similar ticking time-bomb in some other data structure, which might be forgotten when a future minor change to the data file format is made. This would not be caught in normal upgrade tests, which would initialize a database using a previous server version (say, 10.2) and then upgrade to 10.3. To catch something like this, you would have to initialize and populate the database in MySQL 5.1.47 or earlier, with a suitable usage pattern that actually causes some unused to contain suitably unlucky garbage bytes. (On a freshly initialized database they could easily be 0.)

MariaDB Server 10.4 introduced a new file format innodb_checksum_algorithm=full_crc32, and MariaDB Server 10.5 made it the default. Any files that were created when that setting is active are guaranteed to write any unused bytes as zeroes. It also fixes a peculiar design decision that some bytes of the page are not covered by any checksum, and that a page is considered valid if any of the non-full_crc32 checksums happen to produce a match. This includes the magic 0xdeadbeef for innodb_checksum_algorithm=none.

Maybe we should consider eventually deprecating write support for the non-full_crc32 format, to force a fresh start.

Note: We still spend effort to allow in-place binary upgrades. My last example of that was https://jira.mariadb.org/browse/MDEV-29694, removing one of the innovative InnoDB features: the change buffer a.k.a. insert buffer, because it has been the source of many hard-to-debug index corruption bugs.

With best regards,

Marko

On Thu, Dec 8, 2022 at 6:43 PM Reindl Harald h.reindl@thelounge.net wrote:

and when you are at it get rid of "ib_logfile0" and "ib_logfile1" which hold data even in file-per-table mode you can't cleanup from crap caused by a crash 13 years ago

How can that happen when every once in a while upgrades require a full clean shutdown with innodb_fast_shutdown=0 so that the ib_logfile* have to be rebuild due to an on-disk format change? More to the point, if that is a concern, why can you not shut down cleanly with innodb_fast_shutdown=0, rm the files, and let MariaDB re-create them completely clean and empty?

On Thu, Dec 8, 2022 at 6:59 PM Reindl Harald h.reindl@thelounge.net wrote:

Am 08.12.22 um 17:51 schrieb Gordan Bobic:

...

On Thu, Dec 8, 2022 at 6:43 PM Reindl Harald h.reindl@thelounge.net wrote:

...

and when you are at it get rid of "ib_logfile0" and "ib_logfile1" which hold data even in file-per-table mode you can't cleanup from crap caused by a crash 13 years ago

How can that happen when every once in a while upgrades require a full clean shutdown with innodb_fast_shutdown=0 so that the ib_logfile* have to be rebuild due to an on-disk format change?

in my whole life from MySQL 5.0 (the frist time i built mysql with innodb support at all) to MariaDB 10.3 i didn't do anything then upgrade, restart service and be done

I don't have an explanation for that, short of pure luck, in the fact that the log was naturally empty because everything had been flushed to disk at the point you upgraded. You should play the lottery more often if your luck is that good. Mine certainly isn't.

...

More to the point, if that is a concern, why can you not shut down cleanly with innodb_fast_shutdown=0, rm the files, and let MariaDB re-create them completely clean and empty?

don't ask me why that shit can't be removed automatically over 13 years - the reference to './dbmail/#sql2-704-271.ibd' lives somewhere in the "global tablespace" which shouldn't exist at all

I have seen phantom InnoDB tables like that before, but never ones that coiuldn't be cleared with a simple "drop table". Does it show up in innodb_sys_tables or innodb_sys_tablespaces?

Am 09.12.22 um 09:00 schrieb Marko Mäkelä:

Regarding the dbmail/#sql2-704-271.ibd in your other message, in MariaDB 10.5 or later you should be able to drop the table even if a .frm file does not exist, with DROP TABLE dbmail.`#mysql50##sql2-704-271`;

currently on 10.3 because i don't agree the json nosense in the system table.....

If you are using an older version, you could try to copy the .frm file of another InnoDB table to that name

i did that 13 years ago to get rid of that error message, a few years ago 10.1 or 10.2 decided to crash at startup with that non-mathcing file

and then remove the file.

i think you mean "remove the table" - no, "table don't" exist but on the other hand i wasn't able to guess `#mysql50##sql2-704-271` as name and "show tables" don't list anything

hence when you don#t list it in "show tables" just remove it from global tablespace or get rid of that useless global tablewspace at all in "files-per-table" mode

DDL operations should be crash-safe and most of them are atomic starting with MariaDB 10.6.

13 years too late :-)

Am 08.12.22 um 17:47 schrieb Gordan Bobic:

On Thu, Dec 8, 2022 at 6:25 PM Marko Mäkelä marko.makela@mariadb.com wrote:

...

MariaDB Server 10.4 introduced a new file format innodb_checksum_algorithm=full_crc32, and MariaDB Server 10.5 made it the default. Any files that were created when that setting is active are guaranteed to write any unused bytes as zeroes. It also fixes a peculiar design decision that some bytes of the page are not covered by any checksum, and that a page is considered valid if any of the non-full_crc32 checksums happen to produce a match. This includes the magic 0xdeadbeef for innodb_checksum_algorithm=none.

Maybe we should consider eventually deprecating write support for the non-full_crc32 format, to force a fresh start.

Please don't. Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none

that's nonsense - when mariadb writes wrong data in it's files no filesystem can magically fix that

you need to understand what innodb checksums are and then it's logical that the file-system layer is a completly different world

https://dba.stackexchange.com/questions/171708/what-is-an-innodb-page-checks...

and/or have databases many terabytes in size. With terabyte size databases, doing a mysqldump+restore is realistically _never_ going to happen.

that's is the real problem - and hence i expect from software in 2022 that it's able to read it's old data and create new initialized data-files at runtime

Am 08.12.22 um 18:08 schrieb Gordan Bobic:

On Thu, Dec 8, 2022 at 7:02 PM Reindl Harald h.reindl@thelounge.net wrote:

...

...

...

MariaDB Server 10.4 introduced a new file format innodb_checksum_algorithm=full_crc32, and MariaDB Server 10.5 made it the default. Any files that were created when that setting is active are guaranteed to write any unused bytes as zeroes. It also fixes a peculiar design decision that some bytes of the page are not covered by any checksum, and that a page is considered valid if any of the non-full_crc32 checksums happen to produce a match. This includes the magic 0xdeadbeef for innodb_checksum_algorithm=none.

Maybe we should consider eventually deprecating write support for the non-full_crc32 format, to force a fresh start.

Please don't. Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none

that's nonsense - when mariadb writes wrong data in it's files no filesystem can magically fix that

MariaDB can't fix it either. And if that is what happened, there is no benefit to duplicating the effort.

MariaDB does the same as the filesystem InnoDB in fact is more ore less a FS on top of a FS

...

you need to understand what innodb checksums are and then it's logical that the file-system layer is a completly different world

https://dba.stackexchange.com/questions/171708/what-is-an-innodb-page-checks...

You need to understand that properly thought out and sensibly written file systems (which is, granted, pretty rare, I know of a total of 1) implicitly prevent torn pages from being possible. So the checksum and the doublewrite are completely redundant in such cases, and can be safely disabled.

are you dumb or why don't you understand that the filesystem is a completly different layer and has no clue about the data itself?

"Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none" makes you looking like a fool - period

Op 08-12-2022 om 18:59 schreef Gordan Bobic:

On Thu, Dec 8, 2022 at 7:28 PM Reindl Harald h.reindl@thelounge.net wrote:

...

...

...

...

...

MariaDB Server 10.4 introduced a new file format innodb_checksum_algorithm=full_crc32, and MariaDB Server 10.5 made it the default. Any files that were created when that setting is active are guaranteed to write any unused bytes as zeroes. It also fixes a peculiar design decision that some bytes of the page are not covered by any checksum, and that a page is considered valid if any of the non-full_crc32 checksums happen to produce a match. This includes the magic 0xdeadbeef for innodb_checksum_algorithm=none.

Maybe we should consider eventually deprecating write support for the non-full_crc32 format, to force a fresh start.

Please don't. Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none

that's nonsense - when mariadb writes wrong data in it's files no filesystem can magically fix that

MariaDB can't fix it either. And if that is what happened, there is no benefit to duplicating the effort.

MariaDB does the same as the filesystem InnoDB in fact is more ore less a FS on top of a FS

So why do it at both levels? And what makes doing it at MariaDB level in any way better than doing it somewhere else?

...

...

...

you need to understand what innodb checksums are and then it's logical that the file-system layer is a completly different world

https://dba.stackexchange.com/questions/171708/what-is-an-innodb-page-checks...

You need to understand that properly thought out and sensibly written file systems (which is, granted, pretty rare, I know of a total of 1) implicitly prevent torn pages from being possible. So the checksum and the doublewrite are completely redundant in such cases, and can be safely disabled.

"Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none" makes you looking like a fool - period

are you dumb or why don't you understand that the filesystem is a completly different layer and has no clue about the data itself?

Are you too dumb to understand that if a block is corrupted at InnoDB level MariaDB can't do anyting to fix it, but if a block is corrupted at lower level, ZFS can fix it from redundantly stored data and MariaDB never gets to ingest a corrupted block in the first place? If you disagree, please describe a scenario in which an InnoDB page checksum does anything useful if the file system it is on has built in block checksumming and data redundancy.

_______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp

I don't understand much of this discussion, but  take it it's not meant for me...

I take it it does not change anything for the steps I described in my last mail. to get a functioning database again..

regards, Jogchum

On Thu, Dec 8, 2022 at 9:42 PM Reindl Harald h.reindl@thelounge.net wrote:

Am 08.12.22 um 18:59 schrieb Gordan Bobic:

...

On Thu, Dec 8, 2022 at 7:28 PM Reindl Harald h.reindl@thelounge.net wrote:

...

MariaDB does the same as the filesystem InnoDB in fact is more ore less a FS on top of a FS

So why do it at both levels?

because the FS layer can't detect MariaDB errors?

What is the net benefit of detecting said error? The way I see it, the options are: 1) MariaDB detects and error, crashes out 2) MariaDB doesn't detect an error, ingests garbage, crashes out

The only way an error will creep in without the error checking FS spotting it is: 1) manually corrupting the block by writing garbage over it 2) MariaDB generated a duff block and wrote it out 3) Some other hardware failure corrupted the block in MariaDB memory, just before writing it to the file system

If any of the latter set happens, the data is toast anyway. If the former set happens, the data is toast anyway.

Sure it's nice to get an error that confirms that your data is corrupted, but that won't bring it back. Shift it down to a layer where at least a subset of the problemspace can be fixed and we have a net gain in at least some cases.

...

And what makes doing it at MariaDB level in any way better than doing it somewhere else?

which magic should do it somewehre else?

If a file system is in control of data mirroring and checksumming every 16KB block, then if the data is corrupted on disk, file system will detect it on a read and fetch a good copy from an uncorrupted mirror. Application never knows something went wrong, file system replaces the corrupted block on the other disk and everything carries on uninterrupted.

...

...

"Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none" makes you looking like a fool - period

are you dumb or why don't you understand that the filesystem is a completly different layer and has no clue about the data itself?

Are you too dumb to understand that if a block is corrupted at InnoDB level MariaDB can't do anyting to fix it, but if a block is corrupted at lower level, ZFS can fix it from redundantly stored data and MariaDB never gets to ingest a corrupted block in the first place?

it can at least fail early instead work with corrupted data

...

If you disagree, please describe a scenario in which an InnoDB page checksum does anything useful if the file system it is on has built in block checksumming and data redundancy.

INNDOB CHECKSUMS DETECT DATA CORRUPTION WITHIN MARIADB NOT CAUSED BY ANY FILESYSTEM ISSUE AT ALL

the filesystem can't do that with it's block checksumming and data redundancy because there is *nothing wrong* for the view of the FS layer

one is for consistency of the database one is for consistency of the underlying filesystem

two worlds and i simply don't get why people not understanding such basics work in the IT but to top that talking nonsense on mailing-lists

AGAIN: THE FILESYSTEM CAN'T DO ANYTHING BECAUSE IT'S NOT AFFECTED

So your argument is that the page checksum is there to tell you that your data is corrupted because of either: 1) data was corrupted by the database itself, or 2) a superuser overwriting the block on the disk

In either case, you are not getting the data back either way the the database will stop working. So what is the point? I'd rather have the error fixable than have the knowledge that I have an error I can do nothing about.

On Thu, Dec 8, 2022 at 10:59 PM Reindl Harald h.reindl@thelounge.net wrote:

the next reply-all clown.... boy i am subsribed so you can just respond to the list

You sure whine a lot about inconsequential things.

Am 08.12.22 um 21:13 schrieb Gordan Bobic:

...

On Thu, Dec 8, 2022 at 9:42 PM Reindl Harald h.reindl@thelounge.net wrote:

...

Am 08.12.22 um 18:59 schrieb Gordan Bobic:

...

On Thu, Dec 8, 2022 at 7:28 PM Reindl Harald h.reindl@thelounge.net wrote:

...

MariaDB does the same as the filesystem InnoDB in fact is more ore less a FS on top of a FS

So why do it at both levels?

because the FS layer can't detect MariaDB errors?

What is the net benefit of detecting said error? The way I see it, the options are: 1) MariaDB detects and error, crashes out 2) MariaDB doesn't detect an error, ingests garbage, crashes out

silent data corruption

At what layer? If at anything below, up to and including the file system level, ZFS will detect and correct it for you. MariaDB can at best only detect it for you. So catching it at a lower level wins.

...

The only way an error will creep in without the error checking FS spotting it is: 1) manually corrupting the block by writing garbage over it 2) MariaDB generated a duff block and wrote it out 3) Some other hardware failure corrupted the block in MariaDB memory, just before writing it to the file system

If any of the latter set happens, the data is toast anyway. If the former set happens, the data is toast anyway.

silent data corruption

At what layer? If at anything below, up to and including the file system level, ZFS will detect and correct it for you. MariaDB can at best only detect it for you. So catching it at a lower level wins.

...

Sure it's nice to get an error that confirms that your data is corrupted, but that won't bring it back

silent data corruption

...

Shift it down to a layer where at least a subset of the problemspace can be fixed and we have a net gain in at least some cases.

different world

A far better world that you seem have a hard time accepting exists.

...

...

...

And what makes doing it at MariaDB level in any way better than doing it somewhere else?

which magic should do it somewehre else?

If a file system is in control of data mirroring and checksumming every 16KB block, then if the data is corrupted on disk, file system will detect it on a read and fetch a good copy from an uncorrupted mirror.

moron the files are fine from the viewpoint of the filessystem

At what layer? If at anything below, up to and including the file system level, ZFS will detect and correct it for you. MariaDB can at best only detect it for you. So catching it at a lower level wins. I am talking about total cumulative data integrity up to and including what MariaDB reads from disks. What layer in the stack, exactly, are you trying to fight "silent data corruption" you mentioned above?

...

Application never knows something went wrong, file system replaces the corrupted block on the other disk and everything carries on uninterrupted.

moron the files are fine from the viewpoint of the filessystem

...

...

AGAIN: THE FILESYSTEM CAN'T DO ANYTHING BECAUSE IT'S NOT AFFECTED

So your argument is that the page checksum is there to tell you that your data is corrupted because of either: 1) data was corrupted by the database itself, or 2) a superuser overwriting the block on the disk

and it's impressive how long it takes until you mororn understand such basics

And in that case what exactly will the ckecksum do for you? What do you gain from it? The error message that the block is corrupted just before the database crashes?

...

In either case, you are not getting the data back either way the the database will stop working.

yes, but it's a difference if you have silent data corruption detectet months later when clean backups are gone because retention times

...

So what is the point? I'd rather have the error fixable

moron your filesystem can't fix that type of error

that's why "Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none" was an idiots response

a different response would have been "i don't think that i need innodb checksums and want to disable it" but you moron pretended your filesystem does the same

Let's look at the scenarios here, assuming a block gets corrupted: 1) InnoDB has checksums. 1.1) Database doesn't step on the duff block, error doesn't get detected. Until long after your backup retention expired. 1.2) Data corruption happens in memory before the checksum is calculated. Checksum is based on corrupted data and doesn't help you.

2) InnoDB has no checksums. Database steps on the duff block. 2.1) Database crashes during a query. You get the query and tables accessed during a stack trace. You run a check table and it finds a corrupted table. You restore it from backup. 2.2) Database doesn't crash because the damage merely corrupts a single value but the record structure remains sound.

So it is that 2.2) point where the InnoDB checksum gives you anything. It is a vanishingly narrow case. Sufficiently narrow that if you never encountered a relatively common problem with upgrades due to redo log format changes (there have been several over the years), you are orders of magnitude less likely to have encountered this.

If you're concerned about database corruption then you need to start off by having multiple copies of the data available in an online state if you want to recover from corrupt data without doing a backup. This is what ZFS does with RAID - detects which block is corrupt and then uses other data blocks, including parity, to rebuild the corrupt block. MariaDB isn't designed for that. I'm not even sure if there is any database that's designed for that, including SQL Server (see below for more.)

If you want to protect against silent data corruption now there are some alternatives:

1) implement record level checksums that you calculate and verify. This would require adding a "checksum" field to all of your tables. Maybe hide them with stored procedures. This is the best way for you to detect silent data corruption to your data. It won't fix data corruption.

2) implement application logic to always store data in 2+ databases and write an application to verify that all copies of the data are the same (think of this as being the application of a ZFS scrub.) E.g. Using Microsoft Witness Server's with mirror'd SQL databases:

https://learn.microsoft.com/en-us/sql/database-engine/database-mirroring/dat...

The above (1) and (2) won't prevent metadata that the database uses to hold your data from being corrupted. That'd require changes to the page table structures used by InnoDB, etc. On top of that, putting checksums in the page tables wouldn't necessarily detect corruption to your data that happens before it gets included in the page's checksum calculation. If you want to detect corruption of data that is returned over a TCP connection to a remote database then you need record level checksums - see (1) above.

You might even decide to do both (1) and (2) - add checksums to your tables and develop your application to always store (INSERT) all data in 2 different databases (potentially on different servers) and develop the application to query the other database/server if there's a checksum failure when returning data from the first source.

Having multiple databases on different servers also protects you from problems with upgrades, etc, as long as your procedures validate each database is sound before making them "live".

You have to take ownership of the risk for only keeping one copy of critical data. You have to take ownership of the risk for not keeping current backups of your database.

Data resiliency requires effort. If you don't put in any effort then you won't get much in the way of resiliency from corruption. That btw includes good practice for doing ugprades, shutdowns, etc.

MariaDB has as much resiliency to silent data corruption as you put in effort to mitigate it.

On Fri, Dec 9, 2022 at 2:43 AM Reindl Harald h.reindl@thelounge.net wrote:

...

If you're concerned about database corruption then you need to start off by having multiple copies of the data available in an online state if you want to recover from corrupt data without doing a backup.

completly different topic

...

This is what ZFS does with RAID - detects which block is corrupt and then uses other data blocks, including parity, to rebuild the corrupt block.

completly different layers

...

MariaDB isn't designed for that. I'm not even sure if there is any database that's designed for that, including SQL Server (see below for more.)

the topic was "Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none" where you mix two completly independent layers and now we have two morons

Virtually all data corruption that occurs happens in the disk layer, not the database internals layer. This is the reality you seem to be struggling to grasp. If address the problem of corruption in the disk layer, the usefulness of additionally checking in a place that is only useful for internals debugging is vanishingly narrow.

Am 09.12.22 um 08:10 schrieb martin doc:

If you want protection from your database files being corrupted then you need to implement it.

tell me something new - but what has this to do woth the topic?

If you do nothing then you'll have no protection. You did nothing and your data got corrupted.

tell me something new - but what has this to do woth the topic?

There are lots of "best practices" for database administration to deal with this and it appears that you followed none of them.

because i don't disable innodb checksums? how much fog is in your head?

This is nobody's problem but yours - own it.

are you drunken or on drugs?

no other explaination of the brainfuck above as reaction not let "Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none" unchallenged

is this a second account from the same idiot or are you two unique fools - what about marry each other and leave the world in peace?

Am 10.12.22 um 03:42 schrieb martin doc:

...

"Some of us run MariaDB on file systems that do their own block

checksumming, and thus run innodb_checksum_algorithm=none" unchallenged

Your choice. Your risks

wow

it took long until you understood what i was talking about

your offlist pissing contest 4 minutes before this mail was the time you realized that you where a monkey fighting against me while you meant this guy?

that happens when one isn't smart enough to understand who said what and replying unaksed in the middle of a thread

On Fri, Dec 9, 2022 at 12:31 AM Reindl Harald h.reindl@thelounge.net wrote:

...

2.2) Database doesn't crash because the damage merely corrupts a single value but the record structure remains sound.

So it is that 2.2) point where the InnoDB checksum gives you anything

moron it don't matter if you find it useful - the whole point was that you pretended the filesystem can do the same with it's checksums which is nonsense

You are conveniently ignoring the fact that in the vast majority of cases what InnoDB checksums will catch is silent disk corruption rather than database internals corruption. So the one narrow edge case you are clinging to as the full justification of your abusive behaviour and delusions of grandeur are a tiny fraction of a percent of the errors that will cause InnoDB checksums to fail - and outside that narrow edge case all of the rest of them will be caught and handled better at layers below the database itself. So either you are arguing in bad faith, or you really are extensively ignorant of typical failure patterns.

On Fri, Dec 9, 2022 at 2:40 AM Reindl Harald h.reindl@thelounge.net wrote:

...

...

...

2.2) Database doesn't crash because the damage merely corrupts a single value but the record structure remains sound.

So it is that 2.2) point where the InnoDB checksum gives you anything

moron it don't matter if you find it useful - the whole point was that you pretended the filesystem can do the same with it's checksums which is nonsense

You are conveniently ignoring the fact that in the vast majority of cases what InnoDB checksums will catch is silent disk corruption rather than database internals corruption.

i ignore nothing but filesystem corruption is still a different topic

...

So the one narrow edge case you are clinging to as the full justification of your abusive behaviour and delusions of grandeur are a tiny fraction of a percent of the errors that will cause InnoDB checksums to fail - and outside that narrow edge case all of the rest of them will be caught and handled better at layers below the database itself.

the topic was and still is "Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none" where you mix two completly independent layers

...

So either you are arguing in bad faith, or you really are extensively ignorant of typical failure patterns.

the topic was and still is "Some of us run MariaDB on file systems that do their own block checksumming, and thus run innodb_checksum_algorithm=none" where you mix two completly independent layers

They are different layers, but 99.9%+ of corruption that InnoDB checksums ever detects occur in the storage layer, not in the database internals layer. So in terms of the overall picture of the corruption cause probability landscape which you seem to be struggling to see, you are about 0.1% correct. I'll grant you that.

I'll go as far as hazarding a guess that InnoDB checkums were originally added with the main motivation of detecting disk corruption rather than internals debugging. Unfortunately, the code tree back in 3.23.24 (as far as I can tell the release where InnoDB was merged) doesn't seem to contain any annotations on the subject that might shed light on the original motivation.