mwtn35
asked on
Virtual Server keeps shutting itself off
We have a virtual server that has continued to shut itself off periodically for the last week or so. When I go into check the event viewer to see what the error is, it states "The Com+ Event System failed to create an instance of the subscriber", and the Event ID is 4356. Is there anyway to correct this problem to prevent the virtual server from continually shutting itself down?
Thanks
Thanks
Do you mean it's actually shutsdown, and turns off?
Was this created or P2Ved?
ASKER
When I use the technet article, and type in comexp.msc on that virtual server, it says windows can't find the application. Does that mean it's not installed? Any suggestions on how to proceed next?
What OS is it?
ASKER
Yes it actually shuts completely down and I have to login to VMware Vsphere client to restart the virtual machine.
Windows Server 2008?
If you receive the below alert within your event viewer, you’ll find that the id relates to mobsync.exe
To resolve unregister mobsync by typing regsvr32 “%systemroot%\system32\mob sync.dll” /u within the run box and click ok
Event Type: Warning
Event Source: EventSystem
Event Category: Firing Agent
Event ID: 4100
Date: date
Time: time
User: N/A
Computer: computer name
Description: The COM+ Event System failed to create an instance of the subscriber {6295DF2D-35EE-11D1-8707-0 0C04FD9332 7}. CoCreateInstanceEx returned HRESULT 8000401A.
To resolve unregister mobsync by typing regsvr32 “%systemroot%\system32\mob
Event Type: Warning
Event Source: EventSystem
Event Category: Firing Agent
Event ID: 4100
Date: date
Time: time
User: N/A
Computer: computer name
Description: The COM+ Event System failed to create an instance of the subscriber {6295DF2D-35EE-11D1-8707-0
ASKER
Windows Server 2003
Technet Article Applies To: Windows Server 2008
ASKER
Source:Event System
Category:52
Type: Warning
Event ID:4356
User: N/A
Computer: XXXXXX
Description: The COM+ Event System failed to create an instance of the subscriber {6295DF2D-35EE-11D1-8707-0 0C04FD9332 7}. CoCreateInstanceEx returned HRESULT 8000401A.
Category:52
Type: Warning
Event ID:4356
User: N/A
Computer: XXXXXX
Description: The COM+ Event System failed to create an instance of the subscriber {6295DF2D-35EE-11D1-8707-0
Is this a DC or a member server?
ASKER
It's a member server
ASKER
This server is a file and application server as well as print server.
You might try disabling power management features: http://support.microsoft.com/kb/324347
ASKER
Power scheme is Always On
Turn off hard disks- Never
System Standby- Never
Under Advanced
- Prompt for password when computer resumes from standby is checked
- When I press the power button- Shutdown is selected
Hibernation is disabled.
Turn off hard disks- Never
System Standby- Never
Under Advanced
- Prompt for password when computer resumes from standby is checked
- When I press the power button- Shutdown is selected
Hibernation is disabled.
Ok, so in W2K3, the 'components' isn't a snap-in, like in 2K8 per sè. To get to the components, go to Start -> Programs -> Admin Tools -> Component Services. When the MMC opens, drill down the Component Services tree until you get to COM+ Applications. Check the properties of the specific App Component that may be causing the issue. Also, check your COM+ services in the services snap-in.
~coolsport00
~coolsport00
ASKER
Under Services COM+ Event System is set to Automatic and is started, and the COM+ System Applications is set to Manual and is started.
I've drilled down into the Component Services tree and I see nothing wrong under that app component unless I am missing something.
I've drilled down into the Component Services tree and I see nothing wrong under that app component unless I am missing something.
ASKER
Any other ideas how to possibly fix this, still having the issue at least once a day, and normally it's happening in the middle of the night which is getting old for me. :)
One question that comes to mind is do you know that event is the one causing your VM to reboot, or is that all that you noticed in your VM's Event Viewer? You may be pursuing a dead horse there. Outside of what I've provided, I don't have any other answers. Anything else that could potentially be causing the issue? Did you install some update recently that potentially could've triggered this?
ASKER
I know that that is one of the events causing the issue, because it's always right around the time that I get a phone call saying that no one can print, I remote into the VMware box and the server is shut off and I have to turn it back on. There are a couple of SQL Errors prior too, I can post those and see if that sheds any additional light on the problem. And I personally have not installed any new updates on that server.
There are basically two ways (that I know of) that this could be happening.
1. Something in VMware is causing it. This could be someone opening the vSphere client, right clicking the vm, then selecting power off. It could be a script invoking the vSphere CLI or PowerCLI interfaces to initiate a power off of the vm. It could be some other program making a web service call to vCenter (or to the ESX(i) host to initiate a power off. In any event, if this is causing the problem you should be able to select the vm in the client, go to the Tasks & Events tab, and see if there are any Power Off Virtual Machine tasks recorded.
2. Something within the vm is causing it to power off. This is typically done via an ACPI call. To diagnose this consider replaceing the hal and kernel with non-ACPI drivers. To do this open device manager (devmgmt.msc), expand the top Computer node. Typically you will see ACPI Uniprocessor PC or ACPI Multiprocessor PC - right click that select update driver. On the Wizard opening page select "No, not this time" for the question on whether to connect to Windows Update then click Next. Select "Install from a specific location (Advanced)" then click Next. Select "Don't search ..." then click Next. Select "MPS Multiprocessor PC" if you started with "ACPI Multiprocessor PC", or "Standard PC" if you started with "ACPI Uniprocessor PC" then click Next to complete the wizard. After you reboot ACPI functions will be disabled, and if something was calling ACPI to power down the vm the problem should be solved.
Hope this helps...
1. Something in VMware is causing it. This could be someone opening the vSphere client, right clicking the vm, then selecting power off. It could be a script invoking the vSphere CLI or PowerCLI interfaces to initiate a power off of the vm. It could be some other program making a web service call to vCenter (or to the ESX(i) host to initiate a power off. In any event, if this is causing the problem you should be able to select the vm in the client, go to the Tasks & Events tab, and see if there are any Power Off Virtual Machine tasks recorded.
2. Something within the vm is causing it to power off. This is typically done via an ACPI call. To diagnose this consider replaceing the hal and kernel with non-ACPI drivers. To do this open device manager (devmgmt.msc), expand the top Computer node. Typically you will see ACPI Uniprocessor PC or ACPI Multiprocessor PC - right click that select update driver. On the Wizard opening page select "No, not this time" for the question on whether to connect to Windows Update then click Next. Select "Install from a specific location (Advanced)" then click Next. Select "Don't search ..." then click Next. Select "MPS Multiprocessor PC" if you started with "ACPI Multiprocessor PC", or "Standard PC" if you started with "ACPI Uniprocessor PC" then click Next to complete the wizard. After you reboot ACPI functions will be disabled, and if something was calling ACPI to power down the vm the problem should be solved.
Hope this helps...
I guess a third possibility is some logging on and shuting down the PC - but there should be an event log message if that is the case.... So I didn't consider it.
ASKER
Went to the virtual machine itself and looked under the tasks and events like you suggested and here's the error that is recorded when the server shuts itself off.
vcpu-0:ASSERT vmcore/vmm32/platform/comm on/platfor m.c:34 bugNr=17332
vcpu-0:ASSERT vmcore/vmm32/platform/comm
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
I just noticed this resolution refers to an older ESX version. What version of ESX(i) or you running?
Even if it a new version I would still expect it to be NUMA related error.
Even if it a new version I would still expect it to be NUMA related error.
ASKER
It's ESX Server version 3.01 Build 32039.. Is that what you were looking for?
Yup, then the kb article I posted is applicable - follow the steps in the article:
Solution
To resolve this issue:
Check if you are using NUMA architecture. To check if you server is NUMA check for the existence of the /proc/vmware/NUMA/hardware file.
If this file does not exist your server is not NUMA.
In your ESX Server system's advanced configuration options, set the Numa.MonMigEnable value to 0.
To set the Numa.MonMigEnable value to 0:
Choose the host in the VI Client Inventory.
Click the Configuration tab.
Click on the Advanced Options link in the Software box.
A new dialog appears, enter 0 for Numa.MonMigEnable .
Note: You need to follow these steps for all of your NUMA capable ESX hosts
If steps 1 and 2 do not resolve the issue, add the following line to the .vmx file of the virtual machine:
memoverhead = 60
After you complete those steps the problem should be handled.
Solution
To resolve this issue:
Check if you are using NUMA architecture. To check if you server is NUMA check for the existence of the /proc/vmware/NUMA/hardware
If this file does not exist your server is not NUMA.
In your ESX Server system's advanced configuration options, set the Numa.MonMigEnable value to 0.
To set the Numa.MonMigEnable value to 0:
Choose the host in the VI Client Inventory.
Click the Configuration tab.
Click on the Advanced Options link in the Software box.
A new dialog appears, enter 0 for Numa.MonMigEnable .
Note: You need to follow these steps for all of your NUMA capable ESX hosts
If steps 1 and 2 do not resolve the issue, add the following line to the .vmx file of the virtual machine:
memoverhead = 60
After you complete those steps the problem should be handled.
ASKER
The Numa.MonMigEnable value was set to 1 in the advanced settings file, so I have changed it to 0 as they recommmended, so I will see what happens tonight. I appreciate your efforts and willingness to help me out on this!
@bgoering: good catch.
Hopefully this be the issue!
Hopefully this be the issue!
No problem, good luck
Did you check for the existence of the NUMA hardware file?
Did you check for the existence of the NUMA hardware file?
ASKER
Well I was/am going to assume since there were multiple NUMA values set in the advanced settings page that there is, but I am going to go verify for sure.
What is the model of your server?
ditto "bgoering"...I was all outta ideas! :)
ASKER
Since I made the change to that configuration file last night change the NUMA .MonMigEnable from 1 to 0, the server has yet to shut itself off again. I am going to monitor the rest of this week and I will keep you guys updated, in case there is another issue besides this one. Again I appreciate the effort on this, and I would like to keep this open through the first of next week. Thanks
Sounds good to me
ASKER
Great job on assessing my issue, so far this fix seems to be working quite nicely. Thanks!
http://forums.techarena.in/windows-xp-support/409245.htm
http://technet.microsoft.com/en-us/library/cc774410(WS.10).aspx
Regards,
~coolsport00