Host reboots when connecting with client

[solheima]

I just installed ESX 5.1 successfully (to USB). I connect to the web interface as usual and download to client.

However, after installing the vSphere Client, when connecting to the host, the hosts reboots while the client tries to connect (and I can naturally not connect).

Furthermore, I cannot persist changes to the management network via the console. E.g. I want to set the IP configuration to static IP/gateway etc, but the next time the host is booted the configuration is lost.

I can however enable SSH, this change will persist. I can also log on via SSH without problems.

Finally, it seems that if I connect with the client without having made any changes in the management console, I can actually connect and everything seems normal - I can create datastores, VMs...

I have enable only one network card on the server's motherboard. The network card is VmWare Certified (Intel 82574L).

(I can add that I have been running Proxmox on the same physical server for several months without any reboots, so I don't think there is hardware failure)

[spravtek]

Hello, welcome to the community.

There could be a few issues ...

1) Maybe the IP-address you want to use for the host is already in use by something else (IP-Conflict)?

2) USB drive is corrupt/not working properly, did you try with a different USB drive?

[solheima]

Thank you for your suggestions.

The IP adress is not in use by something else, it has been reserved for the same physical box (but when it was not running VmWare) earlier and is reserved for this MAC adress. But I will try to change IP address to something different and set a reservation on this in the router.

I can also try to install to another USB or perhaps to one of the to local SATA SSD (this is my only storage as for now).

[spravtek]

Yes, try testing with a different IP-address ...

And definitely try testing with another USB drive or using the local SSD, there's known cases of strange issues with some USB drives, some USB drives work better than others (better compatibility).

[solheima]

Hi again,

changing the IP address made it possible to persist the changes in the Management Network. However, I still had the same reboot problems.

When skipping the USB and installing directly to one of the local SSDs, things went better. Everything seemed to work normal, I configured the second network adapter / VM Guest Network, uploaded files to data store and created a VM than was running without problems. Unfortunately after working with the host for around 6-7 hours it suddenly rebooted (while not doing anything except an ISO was oploading), Then it rebooted a second time when connecting with the Client.

So it seems there is something with my hardware that ESXi does not like. It is a motherboard with a server-class chipset (Intel C216), and I have disabled all onboard functions not in use (serial, 1394, audio, USB 3.0 etc).

I guess if not anybody has any suggestions I presume it will not be possible to use VmWare on this platform.

[spravtek]

If the system (or all the hardware seperately) aren't in the support matrix then it will be very difficult to get the system running stable ...

You could check the logs, normally you should have logs now it is installed to a fixed disk.

[solheima]

It seems so.

The logs entries in the vSphere Client as I see under events starts at the time it starts to reboot, so I can't see what actually happened that made it reboot. Since the reboot is sudden/immediate, I am not sure if ESXi even manages to collect the error.

Will I see more in the Management Console, if so, which of the logs and what should I look for?

Thanks again,

[spravtek]

This KB: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=203207...

... Should give you info on all the required logs ...

Hostd logs might be interesting ... But be sure to check other logs, they might contain information that is vital ....

[solheima]

Thank you again.

I tried to install another hypervisor (Hyper-V Server 2012) to see if this worked better. This consistently rebooted when I changed network settings, or when I unplugged the network cable. At first I thought there was something wrong with the network cards, but then I tried another thing. The memory on the server is specified for 1600MHz operation, but the motherboard set it to 1333MHz as default. I had manually set it to 1600Mhz (and this server has runned Proxmox for several months without problems), but anyway I returned it to the auto setting in BIOS, which set the memory to run at 1333MHz again. Hyper-V Server 2012 now ran flawless.

I reinstalled ESXi, but unfortunately I still got a reboot, so I guess this is related to my server not being on the compatibility list. I therefore reverted to installing Hyper-V again since this supports most hardware (but I must admit a pain to configure remote management for).

Proxmox is also a level 1 hypervisor, but I guess since it really is an optimized Linux OS (Debian), it handles communication between the memory and the network card in a different way than a more true bare metal hypervisor like ESXi or Hyper-V, and obviously in a way that did not provoke any errors/reboots. Since I did not think of this difference initially I did not try to change these hardware settings.

Anyway, enough rambling from me, but if anyone encounters a similar situation maybe this can help.