- discuss - lists.mariadb.org

[Maria-discuss] Semaphore hangs
by Jon Foster 08 Dec '16

08 Dec '16

We are having trouble with MariaDB hanging due to a "semaphore wait". We then have to shut MariaDB down as it typically won't recover, unless it restarts itself, which happens if we wait long enough. But if its gone on long enough MariaDB won't even shutdown, it hangs indefinitely waiting for some other internal service. I don't remember the exact name and we've been fast enough I haven't seen it in a while. We've had the database on two completely different servers and still see the problem. Both servers were bought new for this project and are a year or less old. They are running all SSD drives, Debian 7 64bit with MariaDB 10.1 from the MariaDB APT repository. Since the XtraDB engine was usually mentioned in the logged messages we switched back to the Oracle InnoDB engine. Although this seems to have reduced the frequency it didn't fix it. Can anyone give some advice on fixing this. It really seams like a bug in MariaDB. I'll try to provide any needed info. TIA - Jon -- Jon Foster jon(a)jfpossibilities.com

2 1

[Maria-discuss] dead server joins the cluster?
by l vic 05 Dec '16

05 Dec '16

Hello, I see the problem with one of the secondary nodes successfully joining the cluster: the cluster_size=3 and it stays this way but database server in container is dead, e.g. nothing listening on the port 3306 Is there possible scenario for the node with failed database server to join the cluster? Thanks,

1 0

[Maria-discuss] New MariaDB releases and EOL distros
by Ian Gilfillan 01 Dec '16

01 Dec '16

Hi From what I understand, 10.2 is currently not being built on Ubuntu 12.04, as after 10.2.2 it required lots of effort to get it working on the platform, too much for a distro about to reach end of life. Currently, it's still built on Centos5, but I understand that this also isn't ideal. The current deprecation dates are listed at: https://mariadb.com/kb/en/mariadb/deprecation-policy/ What about simplifying these decisions with a guideline, such as that major new releases will not be built on distros that reach end-of-life within, say, 6 months of the expected stable release date. So if a new release is expected to be stable in December (understandably always an estimate), we won't bother to build it on distros that EOL before June. Currently, that would mean not building 10.2 on CentOS 5, which I think would be welcomed.

1 0

[Maria-discuss] rejoin timeout
by l vic 01 Dec '16

01 Dec '16

How long can it take for a node to sync during a rejoin? Is there system variable setting that limit? Thanks, -V

1 0

[Maria-discuss] rejoin timeout?
by l vic 30 Nov '16

30 Nov '16

How long can it take for a node to sync during a rejoin? Is there environmental variable for that? Thanks, -V

1 0

Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION)
by Federico Razzoli 30 Nov '16

30 Nov '16

Woops, sorry, I missed the GLOBAL word :) Federico -------------------------------------------- Mer 30/11/16, Ian Gilfillan <ian(a)mariadb.org> ha scritto: Oggetto: Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION) A: maria-discuss(a)lists.launchpad.net Data: Mercoledì 30 novembre 2016, 15:13 On 30/11/2016 15:13, Federico Razzoli wrote: > This always shows the global value: > SHOW VARIABLES LIKE ... > > This always shows the current session value: > SHOW SESSION VARIABLES LIKE ... This is not correct. SHOW VARIABLES shows the session value: SET GLOBAL wait_timeout=601; SET SESSION wait_timeout=599; SHOW SESSION VARIABLES LIKE 'wait_%'; +---------------+-------+ | Variable_name | Value | +---------------+-------+ | wait_timeout | 599 | +---------------+-------+ SHOW GLOBAL VARIABLES LIKE 'wait_%'; +---------------+-------+ | Variable_name | Value | +---------------+-------+ | wait_timeout | 601 | +---------------+-------+ SHOW VARIABLES LIKE 'wait_%'; +---------------+-------+ | Variable_name | Value | +---------------+-------+ | wait_timeout | 599 | +---------------+-------+ _______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss(a)lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp

1 0

Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION)
by Federico Razzoli 30 Nov '16

30 Nov '16

This always shows the global value: SHOW VARIABLES LIKE ... This always shows the current session value: SHOW SESSION VARIABLES LIKE ... Global value is read from the files when the server starts, and later can be changed with SET GLOBAL var = val. Session value is copied from global value when the session starts (so changing the global value doesn't affect existing sessions) and can later be changed with SET var = val. In other words, the global value is important because it is used by all new sessions by default, but the session value is what really takes effect. I beleive that your application sets a session value. Federico -------------------------------------------- Mer 30/11/16, Florent B <florent(a)coppint.com> ha scritto: Oggetto: Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION) A: "Federico Razzoli" <federico_raz(a)yahoo.it>, maria-discuss(a)lists.launchpad.net Data: Mercoledì 30 novembre 2016, 10:44 Ok, I found the problem thanks to you Federico. On Debian, mysql-common package provided by MariaDB repository contains "wait_timeout = 600" in my.cnf. This is not the default value specified in documentation (28800). And when I use "mysql -u root -p -h my_host" command to show variables, the default value is displayed (28800), I never saw 600 in here ! Why ? But displaying variables in my PHP app showed me that wait_timeout was effectively set to 600. Setting wait_timeout=28800 in a config file in /etc/mysql/conf.d directory fixes this. Should we consider "wait_timeout = 600" in my.cnf as a bug ? That's not default value... On 11/29/2016 11:17 PM, Federico Razzoli wrote: > Interesting. A couple ideas: > > 1 > Try this query: > SHOW SESSION VARIABLES LIKE '%timeout%'; > But this time, this should be executed by PHP, using the same library used by the application. Why? To check if a library changes a timeout at session level. > > 2 > Have you checked if your application reconnects N times? For example, 600 could be a 60 seconds timeout * 10 attempts. > > Federico > > > -------------------------------------------- > Mar 29/11/16, Florent B <florent(a)coppint.com> ha scritto: > > Oggetto: Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION) > A: "Reindl Harald" <h.reindl(a)thelounge.net>, maria-discuss(a)lists.launchpad.net > Data: Martedì 29 novembre 2016, 12:22 > > On 11/29/2016 12:03 PM, > Reindl Harald wrote: > > > > > > Am 29.11.2016 um > 11:49 schrieb Florent B: > >> For > example, if I try this in PHP: > >> > >> 1/ Connect to mariadb > >> > >> 2/ sleep 610 > seconds > >> > >> 3/ > ping mariadb > >> > >> => I got Server has gone away > error. > >> > >> > With a 590 sleep time, no error. > >> > >> I use MariaDB 10.1, didn't change > any "timeout" value in MariaDB : > >> interactive_timeout & wait_timeout > are 28800 > > > > http://php.net/manual/en/mysqlnd.config.php#ini.mysqlnd.net-read-timeout > > > > Hi, and > thank you. But I cannot find any timeout of 600 > seconds... > default_socket_timeout = 60 > mysqlnd.net_read_timeout = 31536000 > > And I have the exactly same > problem using this algorithm in Python, it's > not related to PHP... > > _______________________________________________ > Mailing list: https://launchpad.net/~maria-discuss > Post to : maria-discuss(a)lists.launchpad.net > Unsubscribe : https://launchpad.net/~maria-discuss > More help : https://help.launchpad.net/ListHelp > > > _______________________________________________ > Mailing list: https://launchpad.net/~maria-discuss > Post to : maria-discuss(a)lists.launchpad.net > Unsubscribe : https://launchpad.net/~maria-discuss > More help : https://help.launchpad.net/ListHelp

2 1

Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION)
by Federico Razzoli 29 Nov '16

29 Nov '16

Interesting. A couple ideas: 1 Try this query: SHOW SESSION VARIABLES LIKE '%timeout%'; But this time, this should be executed by PHP, using the same library used by the application. Why? To check if a library changes a timeout at session level. 2 Have you checked if your application reconnects N times? For example, 600 could be a 60 seconds timeout * 10 attempts. Federico -------------------------------------------- Mar 29/11/16, Florent B <florent(a)coppint.com> ha scritto: Oggetto: Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION) A: "Reindl Harald" <h.reindl(a)thelounge.net>, maria-discuss(a)lists.launchpad.net Data: Martedì 29 novembre 2016, 12:22 On 11/29/2016 12:03 PM, Reindl Harald wrote: > > > Am 29.11.2016 um 11:49 schrieb Florent B: >> For example, if I try this in PHP: >> >> 1/ Connect to mariadb >> >> 2/ sleep 610 seconds >> >> 3/ ping mariadb >> >> => I got Server has gone away error. >> >> With a 590 sleep time, no error. >> >> I use MariaDB 10.1, didn't change any "timeout" value in MariaDB : >> interactive_timeout & wait_timeout are 28800 > > http://php.net/manual/en/mysqlnd.config.php#ini.mysqlnd.net-read-timeout > Hi, and thank you. But I cannot find any timeout of 600 seconds... default_socket_timeout = 60 mysqlnd.net_read_timeout = 31536000 And I have the exactly same problem using this algorithm in Python, it's not related to PHP... _______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss(a)lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp

1 0

Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION)
by Reindl Harald 29 Nov '16

29 Nov '16

Am 29.11.2016 um 12:22 schrieb Florent B: > On 11/29/2016 12:03 PM, Reindl Harald wrote: >> Am 29.11.2016 um 11:49 schrieb Florent B: >>> For example, if I try this in PHP: >>> >>> 1/ Connect to mariadb >>> 2/ sleep 610 seconds >>> 3/ ping mariadb >>> => I got Server has gone away error. >>> With a 590 sleep time, no error. >>> I use MariaDB 10.1, didn't change any "timeout" value in MariaDB : >>> interactive_timeout & wait_timeout are 28800 >> >> http://php.net/manual/en/mysqlnd.config.php#ini.mysqlnd.net-read-timeout > > Hi, and thank you. But I cannot find any timeout of 600 seconds... > default_socket_timeout = 60 > mysqlnd.net_read_timeout = 31536000 > > And I have the exactly same problem using this algorithm in Python, it's > not related to PHP... then pretty sure you mysqld timeouts are not what you think, otherwise we would see the same issues on our dbmail machines when after hours of idle postfix tries to re-use it's connections (and we did have such events when that values where too low) so why don't you ask the server instead pretend they are not lowered? MariaDB [(none)]> show variables like 'wait_timeout'; +---------------+-------+ | Variable_name | Value | +---------------+-------+ | wait_timeout | 300 | +---------------+-------+ 1 row in set (0.00 sec) MariaDB [(none)]> show variables like 'wait_timeout'; +---------------+--------+ | Variable_name | Value | +---------------+--------+ | wait_timeout | 604800 | +---------------+--------+ 1 row in set (0.00 sec) wait_timeout = 604800 interactive_timeout = 604800

1 0

Re: [Maria-discuss] Aborted connection (CLOSE_CONNECTION)
by Stephane VAROQUI 29 Nov '16

29 Nov '16

Stéphane Varoqui, Senior Consultant Phone: +33 695-926-401, skype: svaroqui http://www.mariadb.com <http://www.mariadb.com/> > Le 29 Nov 2016 à 12:22, Florent B <florent(a)coppint.com> a écrit : > > On 11/29/2016 12:03 PM, Reindl Harald wrote: >> >> >> Am 29.11.2016 um 11:49 schrieb Florent B: >>> For example, if I try this in PHP: >>> >>> 1/ Connect to mariadb >>> >>> 2/ sleep 610 seconds >>> >>> 3/ ping mariadb >>> >>> => I got Server has gone away error. >>> >>> With a 590 sleep time, no error. >>> >>> I use MariaDB 10.1, didn't change any "timeout" value in MariaDB : >>> interactive_timeout & wait_timeout are 28800 >> >> http://php.net/manual/en/mysqlnd.config.php#ini.mysqlnd.net-read-timeout >> > > Hi, and thank you. But I cannot find any timeout of 600 seconds... > default_socket_timeout = 60 > mysqlnd.net_read_timeout = 31536000 > > And I have the exactly same problem using this algorithm in Python, it's > not related to PHP… > Have you check your TCP settings or may be some firewall or proxy that would just sit in the middle ? > _______________________________________________ > Mailing list: https://launchpad.net/~maria-discuss > Post to : maria-discuss(a)lists.launchpad.net > Unsubscribe : https://launchpad.net/~maria-discuss > More help : https://help.launchpad.net/ListHelp

1 0