VmBix 2.0 released

[jackie]

I'm installing vmbix as well. Things seem to be working. I've gotten as far as testing with zabbix_gets. However, I'm also seeing the same warnings on startup as stav13:

14:21:24.672 [main] INFO net.dav3860.VmBix - creating pid file /var/run/vmbix.pid 14974
log4j:WARN No appenders could be found for logger (com.vmware.vim25.ws.XmlGenDom).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
14:21:25.744 [main] INFO net.dav3860.VmBix - starting server on port 12050
14:21:25.744 [main] INFO net.dav3860.VmBix - server started

Is there anyway to get rid of those warnings? The info at the URL didn't mean much to me. Running on Ubuntu14.04.2

Thanks for any info.

[tatapoum]

The VMWare API uses log4j as the default logging architecture and there is no log4j configuration file in your CLASSPATH.
You can get rid of these warnings by :
- creating a /etc/vmbix/log4j.properties file, containing :
log4j.appender.devnull=org.apache.log4j.varia.Null Appender
log4j.rootLogger=fatal, devnull

-adding -Dlog4j.configuration="file:/etc/vmbix/log4j.properties" to the last line in /usr/local/sbin/vmbixd.

[jackie]

That does work, thanks!

Did have to make one change. When I cut and pasted the contents of the log4j.properties file, it didn't work. I had to remove the space between "Null" and "Appender".

[stav13]

Hi,

This is working for me know, dont know how but got the ports all messed up and my accounts kept getting locked. Started afresh, again, and its all working now!

Many Thanks

[tatapoum]

I'm glad that it works now.

[jackie]

We have two vcenters as well, so I'm following the procedure outlined here. Thanks for all the good info!

One question: I would like the hypervisors and vms from each vcenter in different groups on zabbix (e.g. hypervisors_vc1, hypervisors_vc2). Do I understand correctly, that with current templates the hypervisors from both vcenters will end up in the same zabbix group? And similarly the vms?

My thought were that I would duplicate the Template: VmBix vCenter Loadable Module. Then I will customize the Host Prototypes in the Discovery Rules for the ESX host and Virtual Machine, by setting the Host Group to be include the vcenter name (e.g. hypervisors_vc1). Sound reasonable?

Thanks for any feedback.

[tatapoum]

It sounds good to me!

[jackie]

Thanks. Hope you can help me with this issue. We have two vcenters; on one I can run a datastore discovery, the other generates errors and hangs up.

Here's the scenario:

On proxy I manually start vmbix:

# vmbix -P 12050 -u [email protected] -p "xxxxxxx" -s https://xxxx.xxxx.xxxx/sdk
log4j:WARN No appenders could be found for logger (com.vmware.vim25.ws.XmlGenDom).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
14:27:12.647 [main] INFO net.dav3860.VmBix - starting server on port 12050
14:27:12.650 [main] INFO net.dav3860.VmBix - server started

Then from the zabbix server, I can run vmware.ping queries, discover esx hosts, etc. However, when I try to do a datacenter discovery on the zabbix server like:

# zabbix_get -s xxx.xxx.xxx.xxx -p 12050 -k datastore.discovery[*]

On the proxy I see the following:

Exception in thread "Thread-2" java.lang.ClassCastException: com.vmware.vim25.DatastoreInfo cannot be cast to com.vmware.vim25.VmfsDatastoreInfo
at net.dav3860.VmBix$Connection.getDatastores(VmBix.j ava:1673)
at net.dav3860.VmBix$Connection.checkAllPatterns(VmBi x.java:716)
at net.dav3860.VmBix$Connection.run(VmBix.java:3219)

and everything hangs. If I connect to our other vcenter using VMBix, the datastore discovery works great.

Using zabbix 2.4 on Ubuntu 14.04.

Thanks for any ideas.

[tatapoum]

What is the version of your vCenter ?
What kind of storage do you use (SAN, NFS) ?

[jackie]

VMware vCenter Server 6.0.0 build-3018523
A variety of datastores: VMFS5, NFS 3, vsan

[jackie]

Just curious if there is any news on this? Is the variety of datastores we have on our vcenter the source of the problem?

I now have VMBix monitoring two vcenters with loadable modules and things are working great! Did take me some time to modify the loadable module templates for Zabbix to handle two vcenters. I ended up duplicating them, as I wanted the Hypervisor hosts and VMs for each vcenter in different host groups and couldn't see another way to do that.

Anyway, just wanted to say thanks for a great piece of software and all your help with it!

[jackie]

We have been running VMBix with loadable modules for a bit over a week now, and have run into a more serious issue: the memory used on the zabbix proxy is continually increasing. Memory free dropped from 2.5G to 250M. Machine was swapping heavily.

Restarting the zabbix proxy service freed up the memory again, but watching memory available now, and it is continuing to drop.

Is there a way to estimate how much memory the loadable modules requires? Or could it be a memory leak situation?

Would really appreciate some info on this! Thanks!

[tatapoum]

Could you try to test the vmbix loadable module code I uploaded today ?

Can you check with "smem" what is eating your memory on the server ?

[jackie]

I tested the new module code for a few hours now, and continue to see the available memory dropping.

Could the use of strdup in the code be an issue? I believe strdup does a malloc. For example, in the function zbx_module_vmbix, I'm looking at the statement:

zabbix_log(LOG_LEVEL_DEBUG, "Received reply from VmBix. Query: %s, result: %s", strdup(key), strdup(value));

Is that memory allocated by the strdups ever freed? There may be other instances.

Thx.

[Jason]

Is the vmbix project still maintained? Are there any plans to build a module for zabbix 4.0?