VMware Cloud Community
ggerard1
Contributor
Contributor

MySQL 5 not fully auto discovering

Fedora Core 6 on 64bit Intel Xeons

I installed the server and agent (community editions) onto the same machine for a test run.

I see the MySQL process itself. I edited the configuration of it to fill in the jdbc user and password.

On the server logs, I can see all the databases, the tables within and the other performance data that's been gathered. However, no new servers or services are added. I saw something about a 15 minute interval for auto-discovery but it's been longer, and I *see* the data being returned by the plugin on the server log.

Using the hqadmin user.

Looked at:
http://support.hyperic.com/confluence/display/DOCSHQ30/MySQL
0 Kudos
7 Replies
admin
Immortal
Immortal

The servers and services will actually be discovered as separate servers that will not be below the 'MySQL Process' server. If go to Browse Resources then servers do you see your databases?

If not, if you can post the relevant messages from the server.log that could help.
0 Kudos
ggerard1
Contributor
Contributor

Nope, I see tons of EJBs and Postgres stuff from the Hyperic installation, but nothing more than the barest minimum for my MySQL 5.0.22 installation.

Ideas?
0 Kudos
admin
Immortal
Immortal

Oh, are you running HQ 3.0.3? There was a bug in 3.0.2 that caused the detected databases and tables to not be imported correctly. To pick up the fix, you should only need to upgrade the HQ server.

-Ryan
0 Kudos
ggerard1
Contributor
Contributor

Yup, running 3.0.3 everywhere.

As another datapoint, the Mac OS X agent picks up only the barest of information.

One more datapoint, the Linux x64 agent/server doesn't show anything more than the barest of top-style metrics. Here's a copy of my Metrics page. Very few (5) things are available.

What's the root cause for this? I can get most of these from /proc as the user running the agent.

thanks,
greg

Availability LOW AVG PEAK LAST Collection Interval
Availability 100% 00:01:00
System Uptime - - - - NONE
Throughput
Tcp Inbound Connections - - - - NONE
Tcp Outbound Connections - - - - NONE
Tcp State CLOSE - - - - NONE
Tcp State CLOSE_WAIT - - - - NONE
Tcp State CLOSING - - - - NONE
Tcp State ESTABLISHED - - - - NONE
Tcp State FIN_WAIT1 - - - - NONE
Tcp State FIN_WAIT2 - - - - NONE
Tcp State LAST_ACK - - - - NONE
Tcp State LISTEN - - - - NONE
Tcp State SYN_RECV - - - - NONE
Tcp State SYN_SENT - - - - NONE
Tcp State TIME_WAIT - - - - NONE
Utilization
Cpu Idle - - - - NONE
Cpu Idle Time - - - - NONE
Cpu Idle Time per Minute - - - - NONE
Cpu Nice - - - - NONE
Cpu Usage - - - - NONE
Cpu Wait - - - - NONE
Cpu Wait Time - - - - NONE
Cpu Wait Time per Minute - - - - NONE
Free Memory 206 MB 537 MB 1,186 MB 276 MB 00:05:00
Free Memory (+ buffers/cache) 3.4 GB 3.6 GB 3.9 GB 3.5 GB 00:05:00
Load Average 1 Minute - - - - NONE
Load Average 15 Minutes - - - - NONE
Load Average 5 Minutes 0.140 0.294 0.750 0.290 00:05:00
Number of CPUs - - - - NONE
Running Processes - - - - NONE
Swap Free - - - - NONE
Swap Total - - - - NONE
Swap Used 981.6 MB 981.6 MB 981.6 MB 981.6 MB 00:05:00
System Cpu - - - - NONE
System Cpu Time - - - - NONE
System Cpu Time per Minute - - - - NONE
Total Memory - - - - NONE
Used Memory - - - - NONE
Used Memory (- buffers/cache) - - - - NONE
User Cpu - - - - NONE
User Cpu Time - - - - NONE
User Cpu Time per Minute - - - - NONE
0 Kudos
jduino
Enthusiast
Enthusiast

All the metrics listed are the ones it can monitor. However you have to
tell Hyperic to start collecting the data. To do so select the desired
metric(s), then at the bottom of the page enter a collection interval
(typically 1,5, or 10 minutes, but whatever you'd like) then hit the
arrow. After a short time that page will update and you should start
seeing the collected data.

If you want more by default then you need to edit the "Defaults
Configuration" (found under the Administration menu.

John


> One more datapoint, the Linux x64 agent/server doesn't show anything more
> than the barest of top-style metrics. Here's a copy of my Metrics page.
> Very few (5) things are available.
>
> What's the root cause for this? I can get most of these from /proc as the
> user running the agent.
>
> thanks,
> greg
>
> Availability LOW AVG PEAK LAST Collection Interval
> Availability 100% 00:01:00
> System Uptime - - - - NONE
> Throughput
> Tcp Inbound Connections - - - - NONE
> Tcp Outbound Connections - - - - NONE
> Tcp State CLOSE - - - - NONE
> Tcp State CLOSE_WAIT - - - - NONE
> Tcp State CLOSING - - - - NONE
> Tcp State ESTABLISHED - - - - NONE
> Tcp State FIN_WAIT1 - - - - NONE
> Tcp State FIN_WAIT2 - - - - NONE
> Tcp State LAST_ACK - - - - NONE
> Tcp State LISTEN - - - - NONE
> Tcp State SYN_RECV - - - - NONE
> Tcp State SYN_SENT - - - - NONE
> Tcp State TIME_WAIT - - - - NONE
> Utilization
> Cpu Idle - - - - NONE
> Cpu Idle Time - - - - NONE
> Cpu Idle Time per Minute - - - - NONE
> Cpu Nice - - - - NONE
> Cpu Usage - - - - NONE
> Cpu Wait - - - - NONE
> Cpu Wait Time - - - - NONE
> Cpu Wait Time per Minute - - - - NONE
> Free Memory 206 MB 537 MB 1,186 MB 276 MB 00:05:00
> Free Memory (+ buffers/cache) 3.4 GB 3.6 GB 3.9 GB 3.5 GB 00:05:00
> Load Average 1 Minute - - - - NONE
> Load Average 15 Minutes - - - - NONE
> Load Average 5 Minutes 0.140 0.294 0.750 0.290 00:05:00
> Number of CPUs - - - - NONE
> Running Processes - - - - NONE
> Swap Free - - - - NONE
> Swap Total - - - - NONE
> Swap Used 981.6 MB 981.6 MB 981.6 MB 981.6 MB 00:05:00
> System Cpu - - - - NONE
> System Cpu Time - - - - NONE
> System Cpu Time per Minute - - - - NONE
> Total Memory - - - - NONE
> Used Memory - - - - NONE
> Used Memory (- buffers/cache) - - - - NONE
> User Cpu - - - - NONE
> User Cpu Time - - - - NONE
> User Cpu Time per Minute - - - - NONE
>


0 Kudos
ggerard1
Contributor
Contributor

Okay, solved the lack of metrics problem. However, I've gone through many rounds of autodiscovery, changing the password on the JDBC test connection and whatnot but I never get any of my MySQL database instances to show up.

Sorry if I'm being obtuse, but I just don't see what I'm missing.

thanks,
greg
0 Kudos
admin
Immortal
Immortal

Due to all the reports of MySQL discovery being broken, I just gave it a shot with 5.0.37. Seems to work well for me with the exception of some exception handling. If an incorrect username or password is given the UI does not give any indication of error.

I would suggest that you enable agent debug logging on the agent that is monitoring your MySQL instance. Instructions on how to do this can be found here:

http://support.hyperic.com/confluence/display/DOCSHQ30/HQ+Agent+Configuration

After restarting your agent, go back into HQ and resubmit your configuration. You should see the agent try to discover MySQL. If that goes well, check your server log, you should see entries that look like:

2007-04-20 11:35:21,404 INFO [org.hyperic.hq.autoinventory.server.session.RuntimeReportProcessor] Creating new service: localhost.localdomain MySQL 5.x information_schema CHARACTER_SETS

If you are unable to get it working from looking at the logs, attach both your server log and agent log to this forum and I'll have a look.

Thanks Greg,
-Ryan
0 Kudos