[opensuse] MySQL (mariadb) fails to start on 42.3 - Can someone confirm
All, I don't leave the mysql server running continually. I start it when I need to pull in, check databases, etc.. Today, looking at the socket question on this list, I attempted to start mysql and it failed to start. The journal shows: May 08 21:46:41 wizard systemd[1]: Starting MySQL server... May 08 21:46:41 wizard mysql-systemd-helper[3645]: Checking MySQL configuration for obsolete options... May 08 21:46:41 wizard mysql-systemd-helper[3645]: Trying to run upgrade of MySQL databases... May 08 21:46:41 wizard mysql-systemd-helper[3645]: Stale files from previous upgrade detected, cleaned them up May 08 21:46:41 wizard mysql-systemd-helper[3645]: Running protected MySQL... May 08 21:46:41 wizard mysql-systemd-helper[3645]: Waiting for MySQL to start May 08 21:46:41 wizard mysql-systemd-helper[3645]: 190508 21:46:41 [Note] /usr/sbin/mysqld (mysqld 10.0.35-MariaDB) starting as process 3668 ... May 08 21:47:42 wizard mysql-systemd-helper[3645]: MySQL is still dead May 08 21:47:42 wizard mysql-systemd-helper[3645]: MySQL didn't start, can't continue May 08 21:47:42 wizard systemd[1]: mysql.service: Control process exited, code=exited status=1 May 08 21:47:42 wizard systemd[1]: Failed to start MySQL server. May 08 21:47:42 wizard systemd[1]: mysql.service: Unit entered failed state. May 08 21:47:42 wizard systemd[1]: mysql.service: Failed with result 'exit-code'. May 08 21:47:42 wizard sudo[3631]: pam_unix(sudo:session): session closed for user root It has probably been 6 months or more since I started it, so I don't know when it broke. Can someone with 42.3 on their box attempt to start it and let me know if they can confirm this issue? It was previously setup, mysql_secure_installation run and was fully functioning, but now is dead as a doornail. The journal isn't that helpful as to why (other than it just didn't start) I have the following installed, all from OSS/Update: mariadb-10.0.35-35.1.x86_64 mariadb-client-10.0.35-35.1.x86_64 mariadb-errormessages-10.0.35-35.1.x86_64 -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 5/8/19 7:58 PM, David C. Rankin wrote:
All,
I don't leave the mysql server running continually. I start it when I need to pull in, check databases, etc.. Today, looking at the socket question on this list, I attempted to start mysql and it failed to start. The journal shows:
May 08 21:46:41 wizard systemd[1]: Starting MySQL server... May 08 21:46:41 wizard mysql-systemd-helper[3645]: Checking MySQL configuration for obsolete options... May 08 21:46:41 wizard mysql-systemd-helper[3645]: Trying to run upgrade of MySQL databases... May 08 21:46:41 wizard mysql-systemd-helper[3645]: Stale files from previous upgrade detected, cleaned them up May 08 21:46:41 wizard mysql-systemd-helper[3645]: Running protected MySQL... May 08 21:46:41 wizard mysql-systemd-helper[3645]: Waiting for MySQL to start May 08 21:46:41 wizard mysql-systemd-helper[3645]: 190508 21:46:41 [Note] /usr/sbin/mysqld (mysqld 10.0.35-MariaDB) starting as process 3668 ... May 08 21:47:42 wizard mysql-systemd-helper[3645]: MySQL is still dead May 08 21:47:42 wizard mysql-systemd-helper[3645]: MySQL didn't start, can't continue May 08 21:47:42 wizard systemd[1]: mysql.service: Control process exited, code=exited status=1 May 08 21:47:42 wizard systemd[1]: Failed to start MySQL server. May 08 21:47:42 wizard systemd[1]: mysql.service: Unit entered failed state. May 08 21:47:42 wizard systemd[1]: mysql.service: Failed with result 'exit-code'. May 08 21:47:42 wizard sudo[3631]: pam_unix(sudo:session): session closed for user root
It has probably been 6 months or more since I started it, so I don't know when it broke. Can someone with 42.3 on their box attempt to start it and let me know if they can confirm this issue?
It was previously setup, mysql_secure_installation run and was fully functioning, but now is dead as a doornail. The journal isn't that helpful as to why (other than it just didn't start)
I have the following installed, all from OSS/Update:
mariadb-10.0.35-35.1.x86_64 mariadb-client-10.0.35-35.1.x86_64 mariadb-errormessages-10.0.35-35.1.x86_64
David, dunno how much this will help you, but FWIW I have a system running OpenSuSE 42.2 and it is running mariadb - # mysql --version mysql Ver 15.1 Distrib 10.0.32-MariaDB, for Linux (x86_64) using readline 5.1 without the issues you are experiencing. I just restarted the service without a hiccup... It's a little older version than the one you are using.... I will give you a heads up also, I was forced to upgrade to 10.3 of MariaDB because 10.0 - 10.2 are broken handling some complex SQL queries... I found out the hard way and am still struggling with some of the fallout (see my previous posting about phpMyAdmin) HTHs Marc... -- <b>Computers: the final frontier. These are the voyages of the user Marc.<br> His mission: to explore strange new hardware. To seek out new software and new applications.<br> To boldly go where no Marc has gone before!<br></b> -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/08/2019 11:25 PM, Marc Chamberlin wrote:
David, dunno how much this will help you, but FWIW I have a system running OpenSuSE 42.2 and it is running mariadb -
# mysql --version mysql Ver 15.1 Distrib 10.0.32-MariaDB, for Linux (x86_64) using readline 5.1
without the issues you are experiencing. I just restarted the service without a hiccup... It's a little older version than the one you are using....
I will give you a heads up also, I was forced to upgrade to 10.3 of MariaDB because 10.0 - 10.2 are broken handling some complex SQL queries... I found out the hard way and am still struggling with some of the fallout (see my previous posting about phpMyAdmin)
HTHs Marc...
Thanks Marc, I found the reason mysql wouldn't start on 42.3 and it has to do with the startup (probably the 'protected' mode) not respecting options set in /etc/my.cnf, but considering them 'unknown'. There was a change within the past year that broke setting the options: no-auto-rehash prompt=\R:\m\_mysql[\d]>\_ (I have used the same /etc/my.cnf for years) Both are valid Mariadb options that I set in /etc/my.cnf under the [client] group. See: https://mariadb.com/kb/en/library/mysql-command-line-client/#options Strange... -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/08/2019 09:58 PM, David C. Rankin wrote:
All,
<snip>
May 08 21:46:41 wizard mysql-systemd-helper[3645]: Waiting for MySQL to start May 08 21:46:41 wizard mysql-systemd-helper[3645]: 190508 21:46:41 [Note] /usr/sbin/mysqld (mysqld 10.0.35-MariaDB) starting as process 3668 ... May 08 21:47:42 wizard mysql-systemd-helper[3645]: MySQL is still dead May 08 21:47:42 wizard mysql-systemd-helper[3645]: MySQL didn't start, can't continue May 08 21:47:42 wizard systemd[1]: mysql.service: Control process exited, code=exited status=1 May 08 21:47:42 wizard systemd[1]: Failed to start MySQL server. May 08 21:47:42 wizard systemd[1]: mysql.service: Unit entered failed state. May 08 21:47:42 wizard systemd[1]: mysql.service: Failed with result 'exit-code'. May 08 21:47:42 wizard sudo[3631]: pam_unix(sudo:session): session closed for user root
I think I've found the problem, but don't know where it is coming from. In , I have: 190508 21:46:41 [ERROR] /usr/sbin/mysqld: unknown option '--no-auto-rehash' 190508 21:46:41 [ERROR] Aborting So where is '--no-auto-rehash' set? I'll check the usual spots I can think of, but if someone knows where this is being set, I'd appreciate it. -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/08/2019 11:27 PM, David C. Rankin wrote:
I think I've found the problem, but don't know where it is coming from. In , I have:
190508 21:46:41 [ERROR] /usr/sbin/mysqld: unknown option '--no-auto-rehash' 190508 21:46:41 [ERROR] Aborting
So where is '--no-auto-rehash' set? I'll check the usual spots I can think of, but if someone knows where this is being set, I'd appreciate it.
There appears to be a bug in the way 42.3 is parsing /etc/my.cnf. The options causing the startup failure are: [mysql] no-auto-rehash [client] prompt=\R:\m\_mysql[\d]>\_ (which have been present in /etc/my.cnf for years) In fact, the documentation for Mariadb specifically shown exactly the use of no-auto-rehash in its setup under the [mysql] section, see: https://mariadb.com/kb/en/library/configuring-mariadb-with-option-files/ (see "Options" at mid-page and the use of "no-auto-rehash" in my.cnf at bottom of page) prompt is a standard option as well, see: https://mariadb.com/kb/en/library/mysql-command-line-client/ Including either in /etc/my.cnf causes mariadb to fail to start with the messages in /var/tmp/mysql-protected.XyzaBc, e.g. 190508 21:36:24 [ERROR] /usr/sbin/mysqld: unknown option '--no-auto-rehash' 190508 21:36:24 [ERROR] Aborting and similar unknown option '--prompt=\R:\m\_mysql(\d)>\_' Logically this would have something to do with some protection that was added in the past year that may not be behaving as intended. Can someone confirm whether I need to open a bug on this? -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
David C. Rankin wrote:
On 05/08/2019 11:27 PM, David C. Rankin wrote:
I think I've found the problem, but don't know where it is coming from. In , I have:
190508 21:46:41 [ERROR] /usr/sbin/mysqld: unknown option '--no-auto-rehash' 190508 21:46:41 [ERROR] Aborting
So where is '--no-auto-rehash' set? I'll check the usual spots I can think of, but if someone knows where this is being set, I'd appreciate it.
There appears to be a bug in the way 42.3 is parsing /etc/my.cnf. The options causing the startup failure are:
[mysql] no-auto-rehash
[client] prompt=\R:\m\_mysql[\d]>\_
What happens if you shift 'no-auto-rehash' down here instead? I'm not sure, but isn't it a client option (rather than server) ? -- Per Jessen, Zürich (8.1°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/09/2019 01:03 AM, Per Jessen wrote:
[mysql] no-auto-rehash
[client] prompt=\R:\m\_mysql[\d]>\_ What happens if you shift 'no-auto-rehash' down here instead? I'm not sure, but isn't it a client option (rather than server) ?
No help. In the past they were both under [client] [mysqld] ... stuff Now, if I set either under [client] = BAM no start :( -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
David C. Rankin wrote:
On 05/09/2019 01:03 AM, Per Jessen wrote:
[mysql] no-auto-rehash
[client] prompt=\R:\m\_mysql[\d]>\_ What happens if you shift 'no-auto-rehash' down here instead? I'm not sure, but isn't it a client option (rather than server) ?
No help. In the past they were both under
[client] [mysqld]
... stuff
Now, if I set either under [client] = BAM no start :(
I don't have a 42.3 system with a database to test it, but it worked for me on 15.1. -- Per Jessen, Zürich (8.9°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/09/2019 01:36 AM, Per Jessen wrote:
Now, if I set either under [client] = BAM no start :( I don't have a 42.3 system with a database to test it, but it worked for me on 15.1.
Thanks for looking at it. This is probably something specific to 42.3. Caught me completely by surprise. Ironically, if I move the prompt to ~/.my.cnf under [mysql] it works fine -- it's just when they are in /etc/my.cnf that it flags them as unknown -- go figure... :( -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
09.05.2019 9:48, David C. Rankin пишет:
On 05/09/2019 01:36 AM, Per Jessen wrote:
Now, if I set either under [client] = BAM no start :( I don't have a 42.3 system with a database to test it, but it worked for me on 15.1.
Thanks for looking at it. This is probably something specific to 42.3. Caught
I installed mariadb on 42.3, added no-auto-rehash to [mysql] section and it starts just fine. The mysql-sytsemd-helper also does not parse anything (more precisely, it does not use results of parsing anywhere); it simply calls /usr/sbin/mysqld --defaults-file=/etc/my.cnf If something tries to invoke mysqld with other options, you either have non-standard package (not from 42.3) or you modified startup script locally.
me completely by surprise. Ironically, if I move the prompt to ~/.my.cnf under [mysql] it works fine -- it's just when they are in /etc/my.cnf that it flags them as unknown -- go figure... :(
~/.my.cnf is not used during system service startup so you simply deleted this option and it works now. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/09/2019 02:05 AM, Andrei Borzenkov wrote:
I installed mariadb on 42.3, added no-auto-rehash to [mysql] section and it starts just fine. The mysql-sytsemd-helper also does not parse anything (more precisely, it does not use results of parsing anywhere); it simply calls
/usr/sbin/mysqld --defaults-file=/etc/my.cnf
If something tries to invoke mysqld with other options, you either have non-standard package (not from 42.3) or you modified startup script locally.
Thanks Andrei, No startup scripts modified and mariadb is from OSS/update, e.g. mariadb-10.0.35-35.1.x86_64 mariadb-client-10.0.35-35.1.x86_64 mariadb-errormessages-10.0.35-35.1.x86_64 This is my exact /etc/my.cnf I have used forever that is now failing on 42.3: [client] [mysqld] innodb_file_format=Barracuda innodb_file_per_table=ON server-id = 1 no-auto-rehash prompt=\R:\m\_mysql[\d]>\_ sql_mode=NO_ENGINE_SUBSTITUTION,STRICT_TRANS_TABLES [mysqld_multi] mysqld = /usr/bin/mysqld_safe mysqladmin = /usr/bin/mysqladmin log = /var/log/mysqld_multi.log Removing no-auto-rehash and prompt=\R:\m\_mysql[\d]>\_ and all is fine. Placing them under [client] or [mysql] for no-auto-rehash makes no difference. They are still unrecognized. The entire error file is: cat /var/tmp/mysql-protected.KaerMe/log_upgrade_run.err 190508 21:36:24 [Note] InnoDB: innodb_empty_free_list_algorithm has been changed to legacy because of small buffer pool size. In order to use backoff, increase buffer pool at least up to 20MB. 190508 21:36:24 [Note] InnoDB: Using mutexes to ref count buffer pool pages 190508 21:36:24 [Note] InnoDB: The InnoDB memory heap is disabled 190508 21:36:24 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins 190508 21:36:24 [Note] InnoDB: GCC builtin __atomic_thread_fence() is used for memory barrier 190508 21:36:24 [Note] InnoDB: Compressed tables use zlib 1.2.8 190508 21:36:24 [Note] InnoDB: Using Linux native AIO 190508 21:36:24 [Note] InnoDB: Using CPU crc32 instructions 190508 21:36:24 [Note] InnoDB: Initializing buffer pool, size = 128.0M 190508 21:36:24 [Note] InnoDB: Completed initialization of buffer pool 190508 21:36:24 [Note] InnoDB: Highest supported file format is Barracuda. 190508 21:36:24 [Note] InnoDB: 128 rollback segment(s) are active. 190508 21:36:24 [Note] InnoDB: Waiting for purge to start 190508 21:36:24 [Note] InnoDB: Percona XtraDB (http://www.percona.com) 5.6.39-83.1 started; log sequence number 1616727 190508 21:36:24 [ERROR] /usr/sbin/mysqld: unknown option '--no-auto-rehash' 190508 21:36:24 [ERROR] Aborting 190508 21:36:24 [Note] InnoDB: FTS optimize thread exiting. 190508 21:36:24 [Note] InnoDB: Starting shutdown... 190508 21:36:25 [Note] InnoDB: Waiting for page_cleaner to finish flushing of buffer pool 190508 21:36:27 [Note] InnoDB: Shutdown completed; log sequence number 1616737 190508 21:36:27 [Note] /usr/sbin/mysqld: Shutdown complete -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
David C. Rankin wrote:
This is my exact /etc/my.cnf I have used forever that is now failing on 42.3:
[client] [mysqld] innodb_file_format=Barracuda innodb_file_per_table=ON server-id = 1 no-auto-rehash prompt=\R:\m\_mysql[\d]>\_ Removing no-auto-rehash and prompt=\R:\m\_mysql[\d]>\_ and all is fine. Placing them under [client] or [mysql] for no-auto-rehash makes no difference. They are still unrecognized.
David, I've found a desktop machine with 42.3 and mariadb installed - mariadb-10.0.35-35.1.x86_64 FWIW, I can reproduce the problem. With those two options (no-auto-rehash, prompt) under [client], mysql does not start. Remove and it's fine. I can't actually find the error in the log though. -- Per Jessen, Zürich (9.2°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
10.05.2019 8:46, Per Jessen пишет:
David C. Rankin wrote:
This is my exact /etc/my.cnf I have used forever that is now failing on 42.3:
[client] [mysqld] innodb_file_format=Barracuda innodb_file_per_table=ON server-id = 1 no-auto-rehash prompt=\R:\m\_mysql[\d]>\_ Removing no-auto-rehash and prompt=\R:\m\_mysql[\d]>\_ and all is fine. Placing them under [client] or [mysql] for no-auto-rehash makes no difference. They are still unrecognized.
David, I've found a desktop machine with 42.3 and mariadb installed -
mariadb-10.0.35-35.1.x86_64
FWIW, I can reproduce the problem. With those two options (no-auto-rehash, prompt) under [client], mysql does not start.
Get your facts right :) mysqld *DOES* start as is obvious when you check running processes while systemd is waiting for start request to complete. What fails is systemd helper script which calls mysqladmin to check for started mysqld and this *client* fails then: leap423:/home/bor # mysqladmin ping mysqladmin: unknown variable 'no-auto-rehash=true' leap423:/home/bor # Of course, any of clients that are called by script may fail, including those at the "install" and "upgrade" staages, so before mysqld itself is started. There is no "problem" here. no-auto-rehash option is supported by one specific client - mysql - so forcing this option on all clients or on mysqld itself will obviously not work.
Remove and it's fine. I can't actually find the error in the log though.
Because helper script "helpfully" redirects output to /dev/null. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Andrei Borzenkov wrote:
10.05.2019 8:46, Per Jessen пишет:
David C. Rankin wrote:
This is my exact /etc/my.cnf I have used forever that is now failing on 42.3:
[client] [mysqld] innodb_file_format=Barracuda innodb_file_per_table=ON server-id = 1 no-auto-rehash prompt=\R:\m\_mysql[\d]>\_ Removing no-auto-rehash and prompt=\R:\m\_mysql[\d]>\_ and all is fine. Placing them under [client] or [mysql] for no-auto-rehash makes no difference. They are still unrecognized.
David, I've found a desktop machine with 42.3 and mariadb installed -
mariadb-10.0.35-35.1.x86_64
FWIW, I can reproduce the problem. With those two options (no-auto-rehash, prompt) under [client], mysql does not start.
Get your facts right :)
mysqld *DOES* start as is obvious when you check running processes while systemd is waiting for start request to complete. What fails is systemd helper script which calls mysqladmin to check for started mysqld and this *client* fails then:
On this system (office37), mysqld does actually not get started. I guess the helper script doesn't make far enough to actually starting it. office37:~ # time systemctl start mysql Job for mysql.service failed because the control process exited with error code. See "systemctl status mysql.service" and "journalctl -xe" for details. real 1m3.236s user 0m0.006s sys 0m0.002s office37:~ # pidof mysqld (nothing) I agree with your diagnosis, the same problem exists with other clients e.g. mysqldump.
There is no "problem" here. no-auto-rehash option is supported by one specific client - mysql - so forcing this option on all clients or on mysqld itself will obviously not work.
I moved the two options in question to [mysql] - and all is well. I have to wonder why it works on 15.1, but I'm not too worried :-) -- Per Jessen, Zürich (10.2°C) http://www.cloudsuisse.com/ - your owncloud, hosted in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
10.05.2019 3:42, David C. Rankin пишет:
On 05/09/2019 02:05 AM, Andrei Borzenkov wrote:
I installed mariadb on 42.3, added no-auto-rehash to [mysql] section and it starts just fine. The mysql-sytsemd-helper also does not parse anything (more precisely, it does not use results of parsing anywhere); it simply calls
/usr/sbin/mysqld --defaults-file=/etc/my.cnf
If something tries to invoke mysqld with other options, you either have non-standard package (not from 42.3) or you modified startup script locally.
Thanks Andrei,
No startup scripts modified and mariadb is from OSS/update, e.g.
mariadb-10.0.35-35.1.x86_64 mariadb-client-10.0.35-35.1.x86_64 mariadb-errormessages-10.0.35-35.1.x86_64
This is my exact /etc/my.cnf I have used forever that is now failing on 42.3:
[client] [mysqld] innodb_file_format=Barracuda innodb_file_per_table=ON server-id = 1 no-auto-rehash
That's rather different from what you said originally. You said that "no-auto-rehash" option was under [mysql] section, not under [mysqld] section. This this is not supprted mysqld option so error is correct. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 5/10/2019 12:52 AM, Andrei Borzenkov wrote:
That's rather different from what you said originally. You said that "no-auto-rehash" option was under [mysql] section, not under [mysqld] section.
This this is not supprted mysqld option so error is correct.
The problem is -- if you put it under [mysql] it still fails -- same error. If you put prompt under [client] -- still fails. I agree [mysql] is the proper "group" for no-auto-rehash and that [client] is the proper group for the prompt. In the past before the mysql/mariadb fork, both were accepted regardless where in my.cnf they were located (unknown options silently ignored). If I put prompt in [client] *before* the [mysqld] group it will accept the prompt and start. If I put the [client] group below the [mysqld] group it fails to start. Something is amiss. -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
11.05.2019 8:41, David C. Rankin пишет:
On 5/10/2019 12:52 AM, Andrei Borzenkov wrote:
That's rather different from what you said originally. You said that "no-auto-rehash" option was under [mysql] section, not under [mysqld] section.
This this is not supprted mysqld option so error is correct.
The problem is -- if you put it under [mysql] it still fails -- same error.
No, it does not fail here as I already told you. If you want someone to debug it for you, then provide exact content of /etc/my.cnf and /etc/my.cnf.d when it fails with this option under [mysql].
If you put prompt under [client] -- still fails.
May be. I'm really not going to search which option is supported by which individual client for you. At this point you should already know that [client] is read by *every* client which means any client program used in startup script may fail causing startup to fail.
I agree [mysql] is the proper "group" for no-auto-rehash and that [client] is the proper group for the prompt. In the past before the mysql/mariadb fork, both were accepted regardless where in my.cnf they were located (unknown options silently ignored).
And how what had been working in the past is relevant? Your question was why it fails here and now.
If I put prompt in [client] *before* the [mysqld] group it will accept the prompt and start. If I put the [client] group below the [mysqld] group it fails to start. Something is amiss.
Again - if you are expecting someone to debug it for you, provide actual mysql configuration files when startup fails. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 05/11/2019 01:03 AM, Andrei Borzenkov wrote:
The problem is -- if you put it under [mysql] it still fails -- same error. No, it does not fail here as I already told you. If you want someone to debug it for you, then provide exact content of /etc/my.cnf and /etc/my.cnf.d when it fails with this option under [mysql].
I didn't ask for debugging help. I asked for confirmation and whether there were any known changes that could produce the problem. For the record, my.cnf is: # noc /etc/my.cnf [client] prompt=\R:\m\_mysql[\d]>\_ [mysql] [mysqld] innodb_file_format=Barracuda innodb_file_per_table=ON server-id = 1 sql_mode=NO_ENGINE_SUBSTITUTION,STRICT_TRANS_TABLES [mysqld_multi] mysqld = /usr/bin/mysqld_safe mysqladmin = /usr/bin/mysqladmin log = /var/log/mysqld_multi.log The contents of /etc/my.cnf.d/ are # noc /etc/my.cnf.d/default_plugins.cnf [server] # noc /etc/my.cnf.d/error_log.cnf [mysqld] log-error = /var/log/mysql/mysqld.log # noc /etc/my.cnf.d/secure_file_priv.cnf [server] secure_file_priv = /var/lib/mysql-files ('noc' simply output the contents without comment or empty lines) Dunno, If it adding no-auto-rehash under [mysql] Works For You, there is no reason to file a bug, it will simply be closed "WORKS FOR ME". -- David C. Rankin, J.D.,P.E. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
David C. Rankin wrote:
On 05/08/2019 09:58 PM, David C. Rankin wrote:
All,
<snip>
May 08 21:46:41 wizard mysql-systemd-helper[3645]: Waiting for MySQL to start May 08 21:46:41 wizard mysql-systemd-helper[3645]: 190508 21:46:41 [Note] /usr/sbin/mysqld (mysqld 10.0.35-MariaDB) starting as process 3668 ... May 08 21:47:42 wizard mysql-systemd-helper[3645]: MySQL is still dead May 08 21:47:42 wizard mysql-systemd-helper[3645]: MySQL didn't start, can't continue May 08 21:47:42 wizard systemd[1]: mysql.service: Control process exited, code=exited status=1 May 08 21:47:42 wizard systemd[1]: Failed to start MySQL server. May 08 21:47:42 wizard systemd[1]: mysql.service: Unit entered failed state. May 08 21:47:42 wizard systemd[1]: mysql.service: Failed with result 'exit-code'. May 08 21:47:42 wizard sudo[3631]: pam_unix(sudo:session): session closed for user root
I think I've found the problem, but don't know where it is coming from. In , I have:
190508 21:46:41 [ERROR] /usr/sbin/mysqld: unknown option '--no-auto-rehash' 190508 21:46:41 [ERROR] Aborting
So where is '--no-auto-rehash' set? I'll check the usual spots I can think of, but if someone knows where this is being set, I'd appreciate it.
The mysql service unit uses /usr/lib/mysql/mysql-systemd-helper - in my version, I see no such option. "--no-auto-rehash" isn't new, so why it wouldn't be known it's a bit weird. -- Per Jessen, Zürich (7.6°C) http://www.cloudsuisse.com/ - your owncloud, hosted in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (4)
-
Andrei Borzenkov -
David C. Rankin -
Marc Chamberlin -
Per Jessen