Tuesday, September 8, 2009

New Dell MP, some inside information

At a site all the new Dell MPs were loaded and some issues popped up. This blog posting will describe these issues.
  1. Not all Dell servers are properly recognized
    Only 5% of the Dell Servers was recognized properly. All other remaining Dell Servers weren’t shown at all in the View Dell > State View > Servers in the Monitoring Pane. When in the Monitoring Pane the View Discovered Inventory was opened and adjusted to show the Dell Unmanaged Windows Server, all the remaining Dell servers were shown. Also the reason was stated: OMSA needed to be of version 5.3 up to 6.1.

    But that is the case, OMSA version 6.1 is installed and operational on all Dell servers. The OpsMgr event log showed no errors what so ever. The Dell MPs landed properly but somehow didn’t recognize the Dell servers.

    Since all Dell MPs were loaded, I decided to bring this MP suite back to the basics. So I deleted the Detailed MP, DRAC MP and the CMC MP. Also the Information Alerts On MP was loaded. This one had to go as well, since Back-To-Basics is the credo.

    However, the Discovery being used here (Dell Server Discovery), runs every 24 hours. Which is neat but now I wanted to see some quick results, so I made an override which reduced it to 5 minutes. I put that override in a MP of its own so that a simple deletion of it will bring the discovery back to its normal rate, 24 hours.

    Bingo! Soon enough the other Dell servers started popping up in the Console as being Dell Servers and ended up being neatly monitored by the Dell MP. So the special MP containing the override on the frequency of the Discovery could be deleted. 

  2. Specific Dell Tasks run from the Console are using default credentials
    image
    When a Dell server is properly recognized there are 11 tasks available to be launched from the OpsMgr Console. However, many of these tasks are set in such a manner that they use the default credentials (User: Root, Password: Calvin).
    image
    These can not be changed. So when these tasks need to be used in an OMSA environment where the credentials have been changed (which is for many companies a compliance issue) these tasks won’t run since the wrong credentials are being used. The only way to go about it is to disable these tasks and recreate them with the new credentials. However, the credentials are being stored in clear text. Which isn’t neat either…

    These tasks are:
    - Check Power Status
    - Power Reset
    - Power On
    - Turn LED Identification On
    - Turn LED Identification Off
    - Power Off Gracefully
    - Power Cycle
    - Force Power Off

18 comments:

John Bradshaw said...

Thx for this Marnix...I have just imported ONLY the Dell.Connections.HardwareLibrary.mp
(I'm assuming that's all u had left after removing the other MP's)...

I am trying to find the Dell Server Discovery object to override it to 5 mins, but cannot locate it. Can u pls tell me where it is so I can override it as u suggest in order to see what the MP finds?

Thx again,
John Bradshaw

Marnix Wolf said...

Hi John.

You are missing a Dell MP, the 'Dell.WindowsServer.Scalable'.

This one needs to be imported as well. With these two MPs (the one you already loaded and the one I just mentioned) you cover much of the Dell Servers.

This posting of mine describes it in detail: http://thoughtsonopsmgr.blogspot.com/2009/09/new-dell-mp-explained-outside.html, Conclusion section: '...On top of that: start small with only the ‘utility’ Dell.Connections.HardwareLibrary and the MP Dell.WindowsServer.Scalable. Go from there and see what it does and what adjustments are needed...'

Good luck with it John.

Best regards,
Marnix Wolf

HIght said...

I am having the same problem of Dell servers not showing up. I ran through your fix and the server count went from 30 to 80 and then quickly back down to 30. I have about 150 servers that should show up. They are all incorrectly stating that OMSA is incorrect version. Any help would be greatly appreciated.

Thanks

LH

Marnix Wolf said...

Hi LH

Thanks for visiting my blog. Strange is the issue you are experiencing. Normally when a server is discovered, it isn't undiscovered afterwards without some action being taken.

Can you run a report of the Dell Unmanaged servers as I have blogged about, found here: http://thoughtsonopsmgr.blogspot.com/2009/09/running-customized-report-dell.html ?

Perhaps this sheds some light on to the matter. Does this report also contain the servers which have been discovered first and are later 'undiscovered'?

On the problematic servers, is OMSA fully operational? Have you already tried a repair / uninstall-install on one problematic server? Did this solve the issue?

It is better to target one problematic server (which is not production critical) and check it out, setp-by-step. And when changing anything, document it and keep it by one change at a time.

Also change the discovery rate by using an override and target it at only that single server you are testing in order to see some quick results.

Hope this helps.

Best regards,
Marnix Wolf

Anonymous said...

All servers showing up now.....Thankyou.

But the Chassis and Remote Access modules are all unmonitored.....Any ideas how to change them to monitored?

Secondly, I still unsure if the Scalable Edition MP is necessary. If I remove it will I get more detail?
Thx,
John Bradshaw

Marnix Wolf said...

Hi John.

The Scalable MP is needed, check out the diagram in this posting: http://thoughtsonopsmgr.blogspot.com/2009/09/new-dell-mp-explained-outside.html

When the Chassis and Remote Access modules must be monitored, the necessary MPs must be loaded as well.

Best regards,
Marnix Wolf

John Bradshaw said...

Hi Marnix,
Sorry, I was not clear, the Chassis and Remote Access MP's are loaded, but the status is the empty circle, - unmonitored still after 4 days. Is there a way to get that to change?

thx
John Bradshaw

Marnix Wolf said...

Hi John.

Until now I haven't use these MPs. But have you read the guides of these MPs? Are the DRACs and CMCs already present in OpsMgr (these SNMP devices)? It seems to me that the DRACs and CMCs must be present in some kind of way in OpsMgr before the MPs can pick them up.

Best regards,
Marnix Wolf

Marnix Wolf said...

Hi John,

just checked the manual (Dell_Server_Management_Pack_Suite_v4.0_UG.pdf)

Page 31 states: 'Dell Network Devices
To discover Dell DRAC and CMC devices, perform the following steps:
1 Follow steps 1 to 4 as described under "Discovery."
2 Select Advanced Computer Discovery, select Network Devices in the
Computer & Device Types menu, and click Next.
3 Type the Start and End of the IP address range that you want to scan,
type the SNMP Community, and select the SNMP version.
4 Click Discover. The Discovery Progress page appears.
5 On the Select Objects to Manage page, select the devices you want to
manage and click Next.
6 Click Finish.
NOTE: You can override the discovery intervals from the Authoring pane of the
Operations console. To override the discovery intervals for iDRAC 6 modular, modify
the discovery interval period for both the discoveries - DRAC5/iDRAC Discovery
and Dell Integrated Remote Access (Modular) Discovery.'

Have you already done this?

Best regards,
Marnix Wolf

Chuck said...

I too am having the problem of Dell servers not showing as Dell managed. I have reinstalled OM. All versions of OM are at 6.2 however the Discovered Inventory (Dell Unmanged Windows Server) view has “Unable to connect to Dell namespace. Install a sup[ported version of OpenManage Server Administrator (V5.3 or later) displayed. I have removed all the Dell MP’s except the Dell Windows Sever (scalable edition) and Dell Base hardware Library. I have changed the discovery interval to 10 minutes. I have Flushed Health Service and state. I have also spent considerable time on the phone with Dell. I have a few servers that will show as managed for a few hours than disappear but the majority will not ever show as a dell managed. Any help would be appreciated.

Marnix Wolf said...

Hi Chuck,

Unfortunetaly I do not have the solution here as well. I see this problem occuring in ANY environment with Dell servers and SCOM, whether there are 1000+ Dell servers or 10 or less.

Personally i think the discovery script from the Dell MP is flawed so Dell should repair it.

Who knows what the next version of the Dell MP brings?

Compared to the old versions of the Dell MPs the most recent version is already a huge improvement. But still it needs some attention.

Bye,
Marnix Wolf

Dark said...

I have encountered this Dell discovery issue. You can enable logging of the discovery by adding an override to the "Dell Server Discovery" and set the LogLevel to 1. Also set the Interval to 900 (15mins) so you can see the behaviour. A log is created in %windir%\Temp which shows the discovery run. I am getting an intermittant failure when the OMSA version is discovered which means that the 'Unmanaged' value is set and the server ends up in the "Dell Unmanaged" group. the error I see in the log is -

21/10/2010 7:16:10 PM ---- ERROR: Object loop retreival error. ErrNumber:-2147023170Description:
21/10/2010 7:16:10 PM ---- INFO: OMSA URL : OMSA Version :
21/10/2010 7:16:10 PM ---- INFO: UnManaged Server object created. Reason: Unsupported OMSA Version. System: xyz
21/10/2010 7:16:10 PM ---- INFO : Script - Dell Server Discovery : End()

-
I will raise a call to Dell for a fix and let you know

Marnix Wolf said...

Hi Dark.

Thanks so much for the update. Really good work. I have forwarded this issue to someone who works for Dell. Will keep you updated as well.

Cheers,
Marnix

Byty said...

Hello any neews regarding the Dell Discovery?
I have the same issue.

Marnix Wolf said...

Hello Byty.

Thanks for visiting my blog. At the moment I don't have more information about this MP.

Cheers,
Marnix Wolf

DarkDim said...

I am checking v4.1 to see if the discovery still has issues

Adam said...

OpenManage 6.4 fixed all my issues with discovery, I went from only seeing about 50% to 99% of the 500 servers i monitor.

Richard Kinser said...

On the servers I am having trouble with, it seems the dell wmi namespace doesn't exist. When trying to connect to root\cimv2\dell using wbemtest I get the error "invalid namespace". So I guess a reinstall should fix this issue. Just curious what would cause this issue on so many servers (98 in my case). Is there a known issue that the Dell WMI namespace might not get created when you install certain versions of OMSA?