ESX Host Monitoring alert

[Randy Santos]

Randy Santos [http://community.zenoss.org/people/rsantos12184] created the discussion

"ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/69780#69780

--------------------------------------------------------------
I am not only new to the Zenoss world but also CentOS, last time i heavily used linux was during the days of Pine. *rusty*

Anyways, I am running Zenoss Core 4.2.0 SP1 (preloaded vmware image provided by Zenoss) and have installed vmware CLI 4.1 - running ESX4... *going to upgrade to ESXi 5U1 in a couple of months.*

The resxtop test command worked perfectly.

In the document it asks me to cd $ZENHOME/libexec i have no idea where to find it, so i went /usr/libexec thinking it was the same thing.  I believe it is because i am using the vmware image by zenoss instead of doing a fresh install of CentOs.

I also cannot find where $ZENHOME/ZenPacks is located to run the check_esxtop command.

I am receiveing two alerts, one is the SNMP agent saying it cannot make contact.  This confuses me since i dont have any snmp modeler plugins selected, and have also set the zSnmptries to 0

The second alert is

/CMD/Fail No data returned for this command  


| agent | zencommand |
| component | null |
| dedupid | ESX01.YBI.LOCAL||/Cmd/Fail|Physical Cpu|3 |
| eventClass | /Cmd/Fail (http://172.18.102.99:8080/zport/dmd/Events/Cmd/Fail) |
| eventClassKey |
|
| eventClassMapping |
|
| eventGroup |
|
| eventKey | Physical Cpu |
| eventState | New |
| evid | 00505693-6581-bd12-11e2-29cb044f6d18 |
| facility |
|
| message | No data returned for command |
| ntevid |
|
| priority |
|
| severity | 3 |
| summary | No data returned for command |


I am wondering if this is due to the 4.1 cli instead of using a 4.0 cli? Or is it possible the zenpack needing addiotional configuration?  I do see Virtual Machines under components, i do see data under virtual machines > graphs.

Under HOST Graphs though it looks to be missing CPU info, and i believe its the whole physical cpu key erorr above.

Should i also be seeing more under components?  Do i need any other modeler plugins besides zenoss.cmd.esx.esxtop?

Thanks everyone.
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/69780#69780]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Randy Santos]

Randy Santos [http://community.zenoss.org/people/rsantos12184] created the discussion

"Re: ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/69785#69785

--------------------------------------------------------------
Using the CLI 4.0 did not work either, i believe its the esxtop command i am sending to the server.
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/69785#69785]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Doug Syer]

Doug Syer [http://community.zenoss.org/people/dsyer%40nwnit.com] created the discussion

"Re: ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/69829#69829

--------------------------------------------------------------
I have to ask.. Did you look at the firewall setting on your esx host?

You may be picking ul the snmp error from the top level device template.  Zenoss will try to poll snmp if there is an snmp templat assigned regardless of modelling.  There is another zpropoert to disable snmp polling all together, forget thr exact name.
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/69829#69829]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Randy Santos]

Randy Santos [http://community.zenoss.org/people/rsantos12184] created the discussion

"Re: ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/69801#69801

--------------------------------------------------------------
Hey Doug, would resxtop test would've worked if it was a firewall issue?

I actually built a zenoss box from scratch instead of using the vmware image and will be trying it again.

I believe some of the esxtop commands being sent to my esx 4 server are not supported, because i do get some information like virtual machines and what not.

Serves me right for using such an older version, i will be going to 5 esxi shortly. Unfortunately their later model of esxi also do not provide information via snmp.
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/69801#69801]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Doug Syer]

Doug Syer [http://community.zenoss.org/people/dsyer%40nwnit.com] created the discussion

"Re: ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/69835#69835

--------------------------------------------------------------
Im not super familiar with thr non enterprise esx zenpacks but i thought the modeller and basic stats where all polled via snmp and that the esxtop monitor is set up by installing thr perl based vmware cli on the collector.  I thought the cli just ssh'd into thr host to pull stats but im not sure if its that or if the cli uses web services.  The couple times i attempted to play with some esxtop stats i was fustrated by library requirements for the esx cli (not a zenoss issue) and decided not to spend more time on it.

Im assuming esxtop uses zencommand to run, you can try to debug those logs to see what is happening or just temp open up the fw on thr esx host.  I dont think the esxtop is trying to pull any 5.x specific commands or anything that hasnt been around for a while.  I may be wasting your time though.

Sent from my iPhone
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/69835#69835]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Randy Santos]

Randy Santos [http://community.zenoss.org/people/rsantos12184] created the discussion

"Re: ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/69844#69844

--------------------------------------------------------------
Doug thanks for the help, i think i am one step closer, everything seems to be working with the exception of the Physical CPU Data Source

Now this is the code within zenoss

${here/ZenPackManager/packs/ZenPacks.zenoss.EsxTop/path}/libexec/check_esxtop -u "${dev/zCommandUsername}" -p ${dev/zCommandPassword} -s ${dev/manageIp} -d "Physical Cpu" -c _Total

i am trying to figure out this command, is the  CommandUsername and CommandPassword just a user name in the vmware environment; like root?

If so what is the manageIp? The IP of the esx host?

I am trying to find out if the "Physical Cpu" is actually a heading the esx4, perhaps it is named something else?

Note to self, you need the 32bit zlib so the vmware 4 cli 64bit can work ^_^

Thanks Doug
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/69844#69844]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[alphasupremicus]

alphasupremicus [http://community.zenoss.org/people/alphasupremicus] created the discussion

"Re: ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/70747#70747

--------------------------------------------------------------
Hi Randy,

<         'Physical Cpu': '".*%s\(%s\)\\\(.*)\""',
---

         'Physical Cpu': '".*%s\(%s\)\\\(.*)\"',

There were (2) double quotes at the end of the string, i.e. "" followed by a single quote '. Removing the extraneous double quote from fixed the generation of /cmd/fail events and now the Physical CPU Data Source works for me on the command line.

You also may want to check out http://community.zenoss.org/docs/DOC-12909 http://community.zenoss.org/docs/DOC-12909 for an updated version of this check_esxtop script that fixes other problems I was having with it modeling my ESXi 5.0 server. I was getting way too many VMs in the VMs list (and they weren't even VMs but just extra processes).

Anyways, the link above does not fix this extra double quote problem, but if you combine all the changes, you seem to get a good script.
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/70747#70747]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[Randy Santos]

Randy Santos [http://community.zenoss.org/people/rsantos12184] created the discussion

"Re: ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/70770#70770

--------------------------------------------------------------
Are you somd kinf of Zenoss Wizard??  That fixed it wouldnt you know double quotes.  You fixed the two problems i've been dealing with for some time now. Thanks so much

Let me ask, is their much of a benefit to the new esxtop you linked before?

Can the cores esxtop run esx 5.0?
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/70770#70770]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]

[alphasupremicus]

alphasupremicus [http://community.zenoss.org/people/alphasupremicus] created the discussion

"Re: ESX Host Monitoring alert"

To view the discussion, visit: http://community.zenoss.org/message/70785#70785

--------------------------------------------------------------
Well, maybe there's not much benefit for the patched esxtop script I linked to on my prior post if you are NOT running an ESXi 5.x series server, but if you are running ESXi 5.x then you really need that patched esxtop script. The "out of the box" check_esxtop script included in Zenoss 4.2 just plain doesn't work against an ESXi 5.x server as it calls things that are NOT virtual machines, virtual machines in the list of guests. It's really awful. I have like 7 virtual machines on one of my ESXi 5.0 servers and the check_esxtop script modeled my server and said I had 250 virtual machines on it and called processes that were clearly just O/S-level processes, virtual machines.

Anyways, I digress....
--------------------------------------------------------------

Reply to this message by replying to this email -or- go to the discussion on Zenoss Community
[http://community.zenoss.org/message/70785#70785]

Start a new discussion in zenoss-users by email
[discussions-community-forums-zenoss--***@community.zenoss.org] -or- at Zenoss Community
[http://community.zenoss.org/choose-container!input.jspa?contentType=1&containerType=14&container=2003]