FromDual TechFeed (de)

How can I search the following string in a text field: ‘%newline,tabluator,b)%’?

Hello all, on some Linux systems I get the following warning during my MySQL database start-up:

[Warning] Buffered warning: Could not increase number of max_open_files to more than 1024 (request: 8192)
[Warning] Buffered warning: Changed limits: max_connections: 214 (requested 505)
[Warning] Buffered warning: Changed limits: table_cache: 400 (requested 512)

What does it mean and is that something I should care about?

Hello all,

I am getting some MySQL warnings in the error log during the start-up:

[Warning] 'user' entry 'root@master' ignored in --skip-name-resolve mode.
[Warning] 'proxies_priv' entry '@ root@laptop4' ignored in --skip-name-resolve mode.

Please suggest how can I prevent such warnings in the MySQL error log?

In Percona Server we have a STATUS variable indicating roughly possible recovery time:

Combined size of InnoDB log files defines how many changes not reflected in the tablespace we may have where innodb_checkpoint_age shows how much changes we actually have at the current moment, being an actual driving factor of recovery time. If you have very large log files allocated but for your workload innodb_checkpoint_age stays low chances are recovery will be quick.
[ InnoDB crash recovery speed in MySQL 5.6
] …

Table of Contents

• MariaDB Lifetime Support Matrix for MariaDB Releases
• MariaDB Corp Enterprise Support Subscriptions
• MySQL Lifetime Support Matrix for MySQL Releases
• Oracle Corp Enterprise Support Subscriptions for MySQL

Lifetime Support Matrix for MariaDB Releases

Overview

FromDual provides software repositories for their MariaDB/MySQL related software:

• DEB Repository for Debian and Ubuntu
• RPM Repository for Redhat, Oracle Linux, Rocky Linux, AlmaLinux, CentOS and Fedora
• RPM Repository for SLES and OpenSuSE

DEB Repository for Debian and Ubuntu

For Debian and Ubuntu: Create a file /etc/apt/sources.list.d/FromDual.list and add the following information to it:

source /etc/os-release
# ID              : debian or ubuntu
# VERSION_CODENAME: buster bullseye focal jammy …

A test page in English…

dpkg --force-all -i socat_1.7.2.3-1_amd64.deb

I wanted to install MySQL Galera Cluster 5.6 on the new Ubuntu 14.04 with provided packages. It was a bit tricky to make it work.

apt-get update
apt-get install mysql-server   # this is still 5.5
# from universe
apt-get install mysql-server-5.6 mysql-client-5.6 mysql-client-core-5.6

So far so good. Everything works fine as MySQL 5.6.

Now we have to download Galera 5.6 from Coderships website:

wget https://launchpad.net/galera/3.x/25.3.5/+download/galera-25.3.5-amd64.deb
wget …

We found that there were similar problems with table names with a dash/minus in the name (e.g. `table-name`).
This seems to be fixed in MySQL 5.6.9 and 5.7.0. See bugs #55556, #62237 and #64821.
Taxonomy upgrade extras: 

Hello poorx7,

I see, that you are using an old version of mpm: 0.9.2 the current version is 0.9.3 which you can download from here.

0.9.3 has some fixes in the area you are currently struggling... See the Release Notes.

And here it looks like something is wrong with your mpm configuration:

5738:2014-08-06 15:05:51.379 - DBG : /usr/bin/zabbix_sender --zabbix-server --port 10051 --host 'debug' --key FromDual.server.check --value '1' -vv
zabbix_sender [5742]: DEBUG: send value error: gethostbyname() …

Can you repeat this at the command line? This is such a basic problem that I expect that it has nothing to do with MySQL but with your environment/application/set-up or testing method…

An mpm agent run should not be started before the previous has ended. To make this sure we have some internal checks (and a kill). So the interval should be bigger than the duration of a (the longest) run. Typically we run mpm every 10 seconds and that is fine in most of the cases. But our biggest DB is around 250 Gbyte. So I am wondering why your runs take much longer. Should be visible by the timestamps of the log (debug) file where most of the time is spent.

As mentioned earlier (now with a nicer error message) somebody or something is killing the actual mpm agent job. This is typically a next mpm agent job who has to wait too long for the actual one. The question is more: Why is the first one lasting for so long…

It is done: v0.9.3.

The following fixes in the file lib/FromDualMySQLmaster.pm should do the job:

-  $$status_ref{'Binlog_file'} = '';
+  $$status_ref{'Binlog_file'} = 'none';

-  $$status_ref{'Binlog_do_filter'} = '';
-  $$status_ref{'Binlog_ignore_filter'} = '';
+  $$status_ref{'Binlog_do_filter'} = 'none';
+  $$status_ref{'Binlog_ignore_filter'} = 'none';

-      $$status_ref{'Binlog_do_filter'} = $ref->{'Binlog_Do_DB'};
-      $$status_ref{'Binlog_ignore_filter'} = $ref->{'Binlog_Ignore_DB'};
+ …

Thanks, got it:

hostnameX FromDual.MySQL.master.Binlog_ignore_filter 1400661892

I think I see what went wrong. Can you please send me the first 10 lines of the file /tmp/FromDualAgentCache.hostnameX.cache:

shell> head -n 10 /tmp/FromDualAgentCache.hostnameX.cache

I see… The only part in the code where exit code 1 can be reached is the signal TERM trap. Either somebody killed the process or mpm killed a long running former mpm process automatically. How do you start mpm agent?