Feed Aggregator

Hello Matthew,

In the proposed way it does not work. In Zabbix a host name must be unique. What you call db_label is the Zabbix hostname.

We have for our MaaS customers the following constellations:

• Several companies
• Master/Slave set-ups
• Galera Cluster set-ups
• Several MySQL instances (mysqld) per server.

We recommend to use the db_label as follows:

[Company.Server_name.Db_fantasy_name]

In your case you possibly can omit the Company tag. So we have:

[Server_A.Prod]

[Server_B.Failover]

[Server_C.Dev1]

[Server_D.Dev2]

I hope this helps you to reflect your organization good enough?

Regards, Oli

Hi,

In the process of setting up the agent I was modifying the FromDualMySQLagent.conf and setting up the databases as in the documentation:

[your_first_db_name]

Type         = mysqld
Modules      = process mysql myisam innodb

The database name should of course be consistant with a hostname on the Zabbix server. I have gotten all this to work correctly.

I am wondering however, is there the ability to change the database name in the config file to a label and then as an option point it at a database name? Such as:

[db_label]

Database  = your_first_db_name
Type         = mysqld
Modules      = process mysql myisam innodb

The reason I ask is that we have multiple servers with the same database names as they are exact copies of one another because they are either main, failover or test servers.

It may just be the case that I have yet to find the correct documentation.

Thanks in advance for all time and effort put into this question.

Some more information about FromDual:

• The Blogs of FromDual employees.
• FromDual presentations hold at various conferences and events.
• Our FromDual software download site.
• FromDual TV.
• Photos of FromDual events.

Some FromDual tools:

• SQL Beautifier
• Database Health Check
• Performance Tuning Key
• FromDual Consulting Tools
• MariaDB/MySQL Configuration Example

Hello Wills,

No problem. I am lucky we finally found it. Any improvement suggestions are welcome!

Oli

Gr, the issue was definitely with the zabbix server name. It was changed on Apr 5th. I’m sorry to have bothered you.

Hello Wills,

First I would check if the host ‘util-mysql0’ matches EXACTLY the hostname in Zabbix server. That is the typical problem. Are you sure this specific item (FromDual.MySQL.mysql.version) exists under this host?

I still do not get what you mean with active and passive. Can you please elaborate?

Best regards, Oli

Must be Zabbix configuration. I meant the agent configuration, does it matter if it is active or passive? Something changed, but I’m not sure what at this point.

Hello Wills,

What about uploading data of some of the relevant modules (mysql or innodb for example)? Have you changed something except upgrading the mpm agent?

With: Monitoring -> Last data you can see if items are uploaded.

You can simulate this manually as follows:

/usr/bin/zabbix_sender –zabbix-server util-zabbix0 –port 10051 –host ‘util-mysql0’ –key FromDual.MySQL.mysql.version –value ’test version’ -vv

Then you should get something like:

Processed 1 Failed 0 Total 1

If this works then something with your mpm config is wrong. If this fails something with your Zabbix server configuration is wrong.

What do you mean with: “Active vs. Passive?” –> Did you change something in the configuration?

Hi Oli,

Thanks for replying. I did not stop the service, I probably should have before upgrading. However, I did stop the zabbix_agent (killall,) and then restarted it. Here is the log for zabbix_agentd.log: [root@util-mysql0 zabbix]# more zabbix_agentd.log 2880:20120412:104635.151 Starting Zabbix Agent [util-mysql0.arw.pvt]. Zabbix 1.8.10 (revision 24303). 2882:20120412:104635.152 agent #0 started [collector] 2884:20120412:104635.152 agent #2 started [listener] 2885:20120412:104635.152 agent #3 started [listener] 2883:20120412:104635.153 agent #1 started [listener] 2886:20120412:104635.153 agent #4 started [active checks] Here an exerpt of the mpm log in debug: 2891:2012-04-12 10:46:56.905 - DBG : Check connection to zabbix server. 2891:2012-04-12 10:46:56.906 - DBG : /usr/bin/zabbix_sender –zabbix-server util-zabbix0 –port 10051 –host ‘util-mysql0’ –key FromDual.server.check –value ‘1’ -vv zabbix_sender [2895]: DEBUG: answer [{ …

Hi Wills,

This is OK because the key does not exist in any template.

  # This tag does NOT exist in templates!!!
  my $lKey    = 'FromDual.server.check';

Did it stop before you did the upgrade or did it stop when you did the upgrade?

Can you set Debug = 4 in the mpm configuration file and let us know the output? Is the zabbix_agent triggered at all?

Regards, Oli

It was working for a while, and it stopped on April 5th. I upgraded to the latest version, and now when I look at the log, I see this:

[root@util-mysql0 etc]# /usr/bin/zabbix_sender –zabbix-server util-zabbix0 –port 10051 –host ‘util-mysql0’ –key FromDual.server.check –value ‘1’ -vv zabbix_sender zabbix_sender [15866]: DEBUG: answer [{ “response”:“success”, “info”:“Processed 0 Failed 1 Total 1 Seconds spent 0.000016”}] Info from server: “Processed 0 Failed 1 Total 1 Seconds spent 0.000016” sent: 1; skipped: 0; total: 1

Any suggestions?

Table of Contents

• Building the Galera Load Balancer
• Starting the Galera Load Balancer
• Querying the Galera Load Balancer
• Draining nodes with Galera Load Balancer
• Removing nodes from and adding nodes to Galera Load Balancer

When you install a Galera Cluster for MySQL for High Availability (HA) it is not enough to install the Database Cluster to achieve this goal. You also have to make the application aware of this HA functionality. This is typically done with some kind of load balancing mechanism between the database and the application.

We have several possibilities how to make such a load balancing possible:

• We build such a load balancing mechanism directly into the application.
• When we use Java or PHP we can use the failover functionality of the connectors (Connector/J, mysqlnd-ms).
• If we cannot touch the application we can put a load balancing mechanism between the application and the database. This can be done with:
  • a Hardware Load Balancer,
  • a Software Load Balancer (Pen, GLB, LVS, HAProxy, ProxySQL, …

https://github.com/ioggstream/ndb_size.pl/tree/add_ondisk

On April 2^nd 2012 FromDual released the new version v0.9 of its Performance Monitor for MySQL (mpm). The new version can be downloaded from here.

The Performance Monitor for MySQL (mpm) is an agent which is hooked into Zabbix. Zabbix is an integrated Enterprise Monitoring solution which can produce performance graphs and alerting.

The changes in the new release are:

New functionality

• A new server module gathers MySQL database specific server informations. This is especially interesting for the Monitoring as a Service customers.
• You can monitor Galera Cluster for MySQL now. All important items of Galera Cluster for MySQL up to version 2.0 are gathered. The important Triggers and Graphs are available. FromDual Performance Monitor for MySQL becomes your indispensable tool for monitoring Galera Cluster!
• Trigger was added on low open_files_limit

Changed functionality

• Item history was reduced from 90 to 30 days to safe space on disk.
• InnoDB items were added and Graphs improved and cleaned-up.
• MyISAM items were …

FreeBSD 9 seems to have some troubles with MySQL 5.5.20. A customer has moved from MySQL 5.0 on Linux to MySQL 5.5 on FreeBSD 9. He experienced a lot of periodic slow downs on the new, much stronger, system which he has not seen on the old Linux box.

This slow downs were also shown in high CPU system time but we could not see any I/O going on.

When we looked into MySQL we have seen many threads in Opening tables state in the MySQL processlist.

The first idea was to increase table_open_cache to 2048 and later to 4096. This made the Opening tables disappear but then we got a significant amount of threads hanging in Copying to tmp table state in the processlist.

So we suspected that those table are going to disk. But we did not see any I/O (with iostat) and Created_tmp_disk_tables did not change significantly but just Created_tmp_tables.

So I suspect some troubles with Memory allocation on FreeBSD 9 with MySQL.

Then the customer has set table_open_cache = 4!!! And suddenly the problems disappeared. I am a bit …

I understand that you have to focus on the future, no problem. My last comment wasn’t for you. But I’m testing compression in a 5.1 (with InnoDB plugin) now and these tables could help me. I will test that on a 5.5 very soon. Thx Cédric

Hello Cédric,

OK. I see. 5.1 is a bit outdated now… Most of our customers are going to 5.5 now or are already on it. So we have to focus on the future.

Oli

Hello, thx for these informations. I regret that these tables are not available in 5.1 with InnoDB plugin.

Hello Cédric,

What the monitor does is the following:

SELECT SUM(compress_time) AS compress_time, SUM(uncompress_time) AS uncompress_time
  FROM information_schema.INNODB_CMP

The MySQL documentation states in chapter The INFORMATION_SCHEMA INNODB_CMP and INNODB_CMP_RESET Tables:

COMPRESS_TIME: Total time in seconds spent in attempts to compress B-tree pages of the size PAGE_SIZE.
UNCOMPRESS_TIME: Total time in seconds spent in uncompressing B-tree pages of the size PAGE_SIZE.

So we store the delta and can see the seconds spent in compressing or uncompressing pages. And I see that these units do not make much sense here. I will fix this.

If you have suggestions how to monitor these information in a better way we are happy to hear your opinion.

Oli