Missing "add/remove mysql" in Administration uri

[Mike]

After install mysqlcluster with cmon
ls -la repo
-rw-r--r-- 1 root root 5557162 2011-04-12 21:35 cmon-1.1.6-64bit-glibc23-mc70.tar.gz
-rw-r--r-- 1 root root 243596811 2011-02-02 22:20 mysql-cluster-gpl-7.1.10-linux-x86_64-glibc23.tar.gz

1.I'm missing "add mysq"l and "remove mysql" in Administration tab on the Dashboard.
I see only "email notification" and "setup".
So after mysqlcluster installaion using configurator there is no mysql APIs in MySQL Server tab(i see only 127.0.0.1).

2. CMON agent is installed per default with cluster-configurator on the Mgmt and Data Nodes too.
But there is no mysql runing on these servers, so i get errors:

Apr 15 11:09:54 : (WARNING) (stats collection): Failed to connect to mysql server on 127.0.0.1:3306 : Error: Can't connect to MySQL server on '127.0.0.1' (111)
Apr 15 11:09:54 : (ERROR) processmgr: restarting '???u' - failed errno:0, Success

Thanks
Mike

[johan (johan-severalnines)]

Hi Mike,
2) is a known bug. Being fixed.
1) did you install the agents on the mysql servers? You need to set mode=agent in cmon.cnf. You use the same cmon package on every host, it is just the mode that differs. On the cluster control server, you should have mode=controller. Did you install using deploy.sh?

Thanks,
Johan

Sent from my iPad

On 15 apr 2011, at 14:03, Mike <email address hidden> wrote:

> Public bug reported:
>
> After install mysqlcluster with cmon
> ls -la repo
> -rw-r--r-- 1 root root 5557162 2011-04-12 21:35 cmon-1.1.6-64bit-glibc23-mc70.tar.gz
> -rw-r--r-- 1 root root 243596811 2011-02-02 22:20 mysql-cluster-gpl-7.1.10-linux-x86_64-glibc23.tar.gz
>
> 1.I'm missing "add mysq"l and "remove mysql" in Administration tab on the Dashboard.
> I see only "email notification" and "setup".
> So after mysqlcluster installaion using configurator there is no mysql APIs in MySQL Server tab(i see only 127.0.0.1).
>
> 2. CMON agent is installed per default with cluster-configurator on the Mgmt and Data Nodes too.
> But there is no mysql runing on these servers, so i get errors:
>
> Apr 15 11:09:54 : (WARNING) (stats collection): Failed to connect to mysql server on 127.0.0.1:3306 : Error: Can't connect to MySQL server on '127.0.0.1' (111)
> Apr 15 11:09:54 : (ERROR) processmgr: restarting '???u' - failed errno:0, Success
>
> Thanks
> Mike
>
> ** Affects: cmon
> Importance: Undecided
> Status: New
>
> --
> You received this bug notification because you are a member of cmon-dev,
> which is the registrant for cmon.
> https://bugs.launchpad.net/bugs/761678
>
> Title:
> Missing "add/remove mysql" in Administration uri
>
> Status in cmon - MySQL Cluster monitoring:
> New
>
> Bug description:
> After install mysqlcluster with cmon
> ls -la repo
> -rw-r--r-- 1 root root 5557162 2011-04-12 21:35 cmon-1.1.6-64bit-glibc23-mc70.tar.gz
> -rw-r--r-- 1 root root 243596811 2011-02-02 22:20 mysql-cluster-gpl-7.1.10-linux-x86_64-glibc23.tar.gz
>
> 1.I'm missing "add mysq"l and "remove mysql" in Administration tab on the Dashboard.
> I see only "email notification" and "setup".
> So after mysqlcluster installaion using configurator there is no mysql APIs in MySQL Server tab(i see only 127.0.0.1).
>
> 2. CMON agent is installed per default with cluster-configurator on the Mgmt and Data Nodes too.
> But there is no mysql runing on these servers, so i get errors:
>
> Apr 15 11:09:54 : (WARNING) (stats collection): Failed to connect to mysql server on 127.0.0.1:3306 : Error: Can't connect to MySQL server on '127.0.0.1' (111)
> Apr 15 11:09:54 : (ERROR) processmgr: restarting '???u' - failed errno:0, Success
>
> Thanks
> Mike
>

[Mike (one6f)]

Hi Johan,

Thanks for update!

I simply missed "add mysql/remove mysql" on the web page rights in the Administration as it used to be in old version.
I missing the way to add new MySQL APIs to CMON.

First time I've installed using deploy.sh and it was everything fine: on all Nodes I get cmon agent running till I renamed Cluster 'default_cluster_1' to 'SWD'. After reboot (all nodes) I get

root@ludwig:~# /usr/local/cmon/sbin/cmon --config-file=/etc/cmon.cnf
we have a config file! - /etc/cmon.cnf
cluster_id=1
......
......
Apr 18 10:18:56 : (INFO) Registering managed cluster with cluster id=1

mysql error message: Duplicate entry '1' for key 'PRIMARY'Apr 18 10:18:57 : (INFO) A cluster already exists with the same name: default_cluster_1
Apr 18 10:18:57 : (INFO)
Apr 18 10:18:57 : (ERROR) Critical error (mysql error code -1) occured - shutting down

Second time I downloaded the mysqlcluster and cmon .tar.gz, placed it in repo directory and install it with
./install-cluster.sh
./bootstrap.sh
root@ludwig:~/mysqlcluster-71/cluster/scripts/tools# ls ../../repo/
cmon-1.1.6-64bit-glibc23-mc70.tar.gz mysql-cluster-gpl-7.1.10-linux-x86_64-glibc23.tar.gz

The script said it has it installed, but in /var/www/ there is no cmon directory.
Then I have started
 ./install-cmon.sh
This script installed cmon only local.
So i have no agent running on MySQL APIs now.
Something gone wrong.

[johan (johan-severalnines)]

Hi Mike,

The rename problem you describe is a known bug too.

To install cmon manually you should do:

  ./install-cmon.sh<-- will install the controller
  ./install-cmon.sh --mode=agent --host=<hostname> -f<filename> <-- will install the agent.

We removed the add/remove mysql because the agents will report in if they are monitoring a mysql server or not.

We have not had time to document it all yet. Sorry for the inconvience.
These things will be fixed shortly.

BR
johan

On 2011-04-18 13.40, Mike wrote:
> Hi Johan,
>
> Thanks for update!
>
> I simply missed "add mysql/remove mysql" on the web page rights in the Administration as it used to be in old version.
> I missing the way to add new MySQL APIs to CMON.
>
> First time I've installed using deploy.sh and it was everything fine: on
> all Nodes I get cmon agent running till I renamed Cluster
> 'default_cluster_1' to 'SWD'. After reboot (all nodes) I get
>
> root@ludwig:~# /usr/local/cmon/sbin/cmon --config-file=/etc/cmon.cnf
> we have a config file! - /etc/cmon.cnf
> cluster_id=1
> ......
> ......
> Apr 18 10:18:56 : (INFO) Registering managed cluster with cluster id=1
>
> mysql error message: Duplicate entry '1' for key 'PRIMARY'Apr 18 10:18:57 : (INFO) A cluster already exists with the same name: default_cluster_1
> Apr 18 10:18:57 : (INFO)
> Apr 18 10:18:57 : (ERROR) Critical error (mysql error code -1) occured - shutting down
>
>
> Second time I downloaded the mysqlcluster and cmon .tar.gz, placed it in repo directory and install it with
> ./install-cluster.sh
> ./bootstrap.sh
> root@ludwig:~/mysqlcluster-71/cluster/scripts/tools# ls ../../repo/
> cmon-1.1.6-64bit-glibc23-mc70.tar.gz mysql-cluster-gpl-7.1.10-linux-x86_64-glibc23.tar.gz
>
> The script said it has it installed, but in /var/www/ there is no cmon directory.
> Then I have started
> ./install-cmon.sh
> This script installed cmon only local.
> So i have no agent running on MySQL APIs now.
> Something gone wrong.
>

[Mike (one6f)]

Hi Johan,

many thanks for the support!
It worked with
./install-cmon.sh --mode=agent --host=192.168.173.135 -f ../../config/cmon.cnf.agent
./install-cmon.sh --mode=agent --host=192.168.173.138 -f ../../config/cmon.cnf.agent

I would just like to make few comments.
1. -h or --help will overwrite controller if it already exist without asking again:
./install-cmon.sh --help
--mode ommitted - installing 'controller' on 192.168.173.139
cmon-1.1.6-64bit-glibc23-mc70.tar.gz
../../repo/cmon-1.1.6-64bit-glibc23-mc70.tar.gz
192.168.173.139: Executing 'rm -rf /usr/local//cmon*'[ok]
192.168.173.139: Executing .....

2. There are no alarms then one of the mgm nodes is down.

Next is simply understanding question.
3. I've stoped both mgm nodes and rebooted the cmon server (there are all separate PC's).
CMON says:
Apr 19 10:17:45 : (INFO) Trying to connect to management servers...
Apr 19 10:18:00 : (INFO) Could not connect to management server(s). - make sure you have an API slot free in config.ini (http:// www.severalnines.com/cmon/doc/apislot.html ). cmon will continue to work but the aggregation of ndb statistics such as diskdata info will not work. Retrying in 10 seconds.

So far so good.
./start-mgmd.sh -h 192.168.173.133
./show.sh
Connected to Management Server at: 192.168.173.133:1186
ERROR Message: The cluster configuration is not yet confirmed by all defined management servers. This management server is still waiting for node 2 to connect.

Could not get configuration
* 4012: Failed to get configuration
* The cluster configuration is not yet confirmed by all defined management servers. This management server is still waiting for node 2 to connect.

And it was started only when I started the second mgm node after
./start-mgmd.sh -h 192.168.173.136
tail -f /var/log/cmon.log
....
Apr 19 10:20:47 : (INFO) Could not connect to management server(s). - make sure you have an API slot free in config.ini (http:// www.severalnines.com/cmon/doc/apislot.html ). cmon will continue to work but the aggregation of ndb statistics such as diskdata info will not work. Retrying in 10 seconds.
Apr 19 10:20:57 : (INFO) Trying to connect to management servers...

Apr 19 10:21:09 : (INFO) Connection ok..
Apr 19 10:21:09 : (INFO)
Apr 19 10:21:09 : (INFO) Please wait while cmon is starting up..

I believed that it enough to have one mgm node to start, or it is a config question?

Thanks
Mike