Currently running 4.1 and of course 4.1 U1 is now available. I don't see many posts on it in here. Is that because not many people have taken the plunge or because it's "just another update" in vsphere terms i.e. not to be considered quite so major as, say, a Windows Service Pack?
Thanks.
You're right, using HDDs instead of flash drives aren't what I've planned...
I have a Dell R710 and a Fujitsu Primergy RX300 S5. Both have exactly the same problem (scratch partition disappears at some point in time).
The Dell has the "original" ESXi pre-install on a flash drive, and I installed the latest 4.1.0.U1 ESXi download from vmware.com onto the flash drive of the Dell 2 weeks ago.
So, not only IBMs are affected, and not only servers installed with the download from VMware are affected.
When I spoke with technicians from Fujitsu and VMware one year ago, they told me that there's no real difference between "embedded" and "installable" versions of ESXi. Some hardware manufacturers only incorporate some drivers into the ESXi setup image to make it more compatible with special devices. But everyone I spoke with was confident that installing the "raw" ESXi download version from vmware.com onto a USB stick and booting this on any server would run without problems.
Dell differenciates between installable and embedded versions and offers downloads of customized versions at their site. Fujitsu has a customized version as well, but only one image without a hint whether it is installable or embedded.
My concerns were that the installable version uses a page file that would corrupt my flash drive...
Andreas
If you aren't adverse to reinstalling the host, why not install the Installable edition on the flash card (or USB stick) instead? Installable doesn't mean you need a hard drive.
That's what I did with the Dell blades having the problem - we just reinstalled right onto the SD card.
I now have all iso images of all available versions (ESXi from VMware, both installable and embedded customized from Dell, and an older 4.1.0 without SP1 from Fujitsu).
I already installed the VMware version onto the Dell. What I could do next is to install the Dell customized embedded version, what I'll probably try next week.
Fujitsu will release 4.1.0.U1 sometime next month, so the Primergy has to wait some more. In the meantime I can look if the customized image for the Dell would make a difference...
But anyway, having physical hard drives as an alternative gives me a slightly better feeling - I know I have hardware to try plan B...
It's sad to see VMware delegating "embedded" questions to hardware manufacturer support departments... What would VMware's comment be if I'd run ESXi on SSDs, a "mixture" between standard HDDs and flash drives?!
Andreas
Edited due to a typo.
What sort of NIC issue are you having with the IBM x3690s? I'm working with a customer with 4 (all embedded) and on 2 we've had the following issue:
Several hosts (IBM x3690 x5) are
1) loosing connectivity to the boot flash device (Hypervisor1 - 3)
2) generating an error on vmnic0. - this is the error - Uplink vmnic0 has recovered from a transient failure due to watchdog timeout
When this happens vmnic0 is not failed over (vSwitch0 has vmnic0 and 4). Rather VMs and the management port continue to use it and become disconnected from the network. Here's a port of the events for the host -
Uplink vmnic0 has recovered from a transient failure due to watchdog timeout Warning 3/17/2011 8:53:54 PM esxi-it2.network.com
Lost connectivity to storage device mpx.vmhba32:C0:T0:L0. Path vmhba32:C0:T0:L0 is down. Affected datastores: "Hypervisor1", "Hypervisor2", "Hypervisor3". error 3/17/2011 8:53:29 PM esxi-it2.network.com
User root logged out info 3/17/2011 8:53:29 PM root
Boot partition /bootbank cannot be found (15:09:20:42.937
my issues are with 10G CNA's and dvSwitches dropping all connectivity after a period of time, this period is interimittent and can occur in 20 minutes or 7 days. the etherchannel is simply killing off one leg and then all outbound vm traffic ceases, inbound traffic is ok, switch the physical nic cables and it reverses, traffic outbound works, inbound ceases. I have a thread about it here: http://communities.vmware.com/message/1721672
I know there is an issue with esxi 4.1 and etherchannel on the vswitch0, there is a KB about it: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=102275...
I have embedded HP for my blades and getting the Error Code : 10.
Reading though this thread I logged into each of my hosts and did the following commands
cd /var
mv tmp tmp.old
mkdir tmp
I tried doing a symlink for /var/tmp to point to /tmp but it still failed.
This worked like a charm.
Edit : Though recreating the directories worked "kinda" ( not sure what else it affected ), ultimately I reinstalled the ESXi server, which resolved the issue.
Thought I would post this link as it relates to this thread also
http://communities.vmware.com/message/1727986#1727986