Hello,
And yes, via the GUI of a vCenter object running version 8.0U2 you cannot manually set a MAC Adddress other than the automatically generated one.
Regarads,
Ferdinando
Hello
This problem did not exist in the previous version, but?
From where ?
So what to do with ovh servers?
Friends, no one should update VMware vCenter Server 8.0.2.00000, there are hundreds of errors.
Hello Sami19861987,
It's a good question, but one that should be asked to the software manufacturer.
But if we want the question is a little more messy, in the sense that in versions prior to ESXi / vCenter 8.0U2 what from the point of view of a vCenter object is a MAC address whose assignment is as "automatic" from the point of view of the GUI of a HOST running ESXi it's as "assigned" MAC Address NOT an "automatic" one but with this latest recent version things have changed in such a way that it is not at all difficult to cause a series of other messes and consequent problems, that is, it is no longer so true.
As you have seen via the GUI of a vCenter object version 8.0U2, the option to assign a MAC Address manually or vice versa does not work. So you might think about getting around the problem by intervening from the GUI of the HOST on which that virtual machine resides, IMHO DO NOT DO THIS if you plan on manually returning to an automatic assignment later, at least NOT with a HOST running ESXi 8.0U2.
I just noted that a standalone HOST running ESXi version 8.0U2 when creating a VM automatically assigns by default a so-called "impermissible static ethernet address 00:0c:29:xx:yy:zz" which, when that host will be subsequently added to inventory of a vCenter object versione 8.0U2 the circumstance is verified and results in an impediment to start/restart of the affected virtual machine.
I can't say if there are other combinations potentially capable of producing other similar results, but I was able to easily reproduce some of them using what the GUIs allow. If there was a channel to report potential BUGS or product defects, it can happen (and I understand it), I would have used those instead of the forum.
Regards,
Ferdinando
Note to moderators: If you find this post "inappropriate", remove it, but please report it to someone who can conduct further investigation.
For example, there are dozens of errors, not just on Mac.
If you watch the other simplest video, the settings made in the GUI are lost when you refresh the page with F5.
Should I build it from scratch? Will this problem persist?
Hello,
I can understand your frustration, but I'm just as much a user of the product as you are, nothing more, nothing less.
Try contacting technical support by opening a ticket and see what they tell you.
I can only tell you that I have built more than one test environment and I have run into the same problem as you (and some other different inconveniences), but I haven't thought about how to get around them because as a result I will keep my (very) small IT infrastructure for a while longer at the previous version 8.0U1c level.
Regards,
Ferdinando
if you have an SR please let us know, I just tested the process in my lab and I hit the same problem indeed.
I filed a bug for it, and am now uploading my logs.
Dear Mr. Depping, good morning,
I don't have a "Service Request" partly because the VMUG licenses do not provide support and in any case even if I wanted to associate it with my customer account I would have to purchase a ticket first. But the point is that I fear it solves little or nothing without a product update.
You'll excuse me, but I think I should edit this post.
Out of my curiosity, I deployed a new host that runs ESXi version 8.0U2 with two virtual machines, both of which received an automatic MAC address in the "00:0c:29:xx:yy:zz" range, first starting them both then turning off one of them and then I added said host to the inventory of a version 8.0U2 vCenter object. Let's get to the practical result, the first (powered on) kept its MAC Address in the "00:0c:29:xx:yy:zz" range as "automatic", the second (powered off) was automatically reconfigured with a MAC Address in the "00:50:56:xx:yy:zz" range as "assigned", then after turning off all the virtual machines and restarting the said host, things changed once again, the previously powered on VM received a new MAC address assigned in the "00:50:56:xx:yy:zz" range as "assigned".
Let's put it in other words, the moment an ESXi object assigns impermissible MAC addresses to my virtual machines, and "I can't do anything about it", then the moment I add a vCenter object to the equation I don't want them modified automatically (I would say to the ambush) to remedy the initial inconsistency. It could potentially mess things up.
Regards,
Ferdinando
How did this make it past IA to GA ? Have they fired the QA department? What a mess this thing is. I have seen countless errors in my test environment.
It's giving me the bad vibes like the 7.0.2 upgrade that destroyed over a dozen SD cards for me.
Such a shame that they keep releasing horribly bugged software that is designed to be run in prod and to be extremely stable.
Guess we'll wait for the U2a or b release before trying again
Hello,
Please excuse me for this long post which I hope is sufficiently clear, I find it a little difficult to express myself in a language other than my own.
In truth, after my last post I continued to go around it, and some of the things I found at that moment can still change depending on other different circumstances that (certainly my limitation) I cannot understand. Let's say that a good starting point is this slightly old but good thread all the same:
https://communities.vmware.com/t5/ESXi-Discussions/Static-MAC-address-not-starting-with-00-50-56/td-...
The essence is that an automatically generated MAC Address in the "00:0C:29.xx.yy.zz" range is implicitly valid thus setting one manually in this same range, should still be valid regardless of any other mechanism today "a problem", because now "reserved" thus "impermissible". When a vCenter version 8.0U2 object finds a virtual machine(s) in its inventory with such a manual MAC address it prevents them from starting, and since it is not possible to manually set a different one, the remedy becomes more complicated than that it should be.
Practical result:
"00:0c:29:27:e7:11 is not an allowed static Ethernet address. It conflicts with VMware reserved MACs. Failed to start the virtual machine. Module DevicePowerOn power on failed. Could not set up "macAddress" for ethernet0. Invalid MAC address specified".
I also noticed that a host that run ESXi version 8.0U2 if you try to manually set a MAC Address within the aforementioned range produces an error saying that it cannot be set because it is "impermissible", but then get sets anyway and the affected VM can no longer start as well.
Practical result:
"Failed to power on virtual machine VM-test. Impermissible static Ethernet address: '00:0c:29:2f:45:9c'. It conflicts with VMware reserved MACs. Click here for more details. dismiss".
With this latest version, when creating a new virtual machine according to the selected GUEST operating system you find yourself with a default USB controller, in version 3.2 from the point of view of a vCenter object and 3.1 from the point of view of an ESXi host, no choice, it must be removed and then added back at which point then you can choose a USB controller version 2.0 or 3.x (as before).
Another thing that worries me quite a bit, what is the expected default behavior for a virtual machine (whether it is a new or pre-existing one) with a hardware level lower than 20 in terms of "chipset.motherboardLayout", I noticed that either after a vMotion operation is performed or because default behavior, with version 8.0 and later I always found this parameter added but it is not clear to me according to "what kind of logic" and whether it is then honored even in the case VHW lower then 20 or not. From my very personal point of view it is not a question of "goat wool" at all.
Practical result:
virtualHW.version = "19"
chipset.motherboardLayout = "acpi"
But if it were just that, one way or another I'll make do anyway.
After the clean deployment of a vCenter 8.0U2 object according to the "Tiny" model I found myself with another set of hassles, after no more than two hours I received the "usual" invitation to add more RAM memory (which I'm now used to ignore) and dozens of events of this kind as warning (never seen before, currently running vCenter object version 8.0U1c):
"Privilege check failed for user VSPHERE.LOCAL\vmware-vsm-37799e0e-cac0-43f7-adde-610f5f1db676 for missing permission Sessions.TerminateSession. Session user performing the check: VSPHERE.LOCAL\vmware-vsm-37799e0e-cac0-43f7-adde-610f5f1db676", "com.vmware.vc.authorization.NoPermission".
Frankly, I don't understand if it's just a cosmetic issue or an indication of a concrete potential problem.
Regards,
Ferdinando
Note for the moderator(s): If you find this post excessively polemical, please remove it. Nonetheless, I agree with the comment of @pmichelli and the OP.
I cannot repro the issue you mention in my lab unfortunately when it comes to the impermissible ranges, tried it on 8.0 U2 just now, but works as expected fortunately/unfortunately, not sure which one.
Hello,
Once again, thank you for your interest,
It is rather difficult to show a series of complex events, writing and/or showing photographs in the context of a forum, also because the operating context can be different, it doesn't take much to highlight unexpected things or behaviors that may even be very different and in any case it could take a long time.
If anyone wanted I could show it to them, which is sometimes worth more than many words.
Regards,
Ferdinando
Hello,
In any case I mean what is reported in the attached file in ".pdf" format and the ".vmx" file (had to rename in .txt) of a virtual machine with "VHW 19" supposed to have no knowledge of a chipset.motherboardLayout = "acpi" parameter meaning.
Regards,
Ferdinando
We were able to find a workaround with following article:
https://kb.vmware.com/s/article/1017348#:~:text=To%20resolve%20this%20issue%2C%20ensure,Port%20Group...
We opened a case and are trying to get a bug fix.
VMware vCenter Server 8.0.2 is broken as hell... not just MAC addresses, but also user without admin permissions is not able to edit settings and connect ISO, or it's not possible to do storage vmotion or import OVA... vCenter from version 7 is **bleep** product, not usable, always broken, unsecure, etc... that's how it looks when VMware started saving money and delegated all work to India...
@zoomprofile wrote:
VMware vCenter Server 8.0.2 is broken as hell... not just MAC addresses, but also user without admin permissions is not able to edit settings and connect ISO, or it's not possible to do storage vmotion or import OVA... vCenter from version 7 is **bleep** product, not usable, always broken, unsecure, etc... that's how it looks when VMware started saving money and delegated all work to India...
Did you file SRs for these issues?
@depping wrote:
@zoomprofile wrote:VMware vCenter Server 8.0.2 is broken as hell... not just MAC addresses, but also user without admin permissions is not able to edit settings and connect ISO, or it's not possible to do storage vmotion or import OVA... vCenter from version 7 is **bleep** product, not usable, always broken, unsecure, etc... that's how it looks when VMware started saving money and delegated all work to India...
Did you file SRs for these issues?
VMware-VCSA-all-8.0.2-22385739.iso cannot be installed at all, and I even doubt whether they released it after testing.
If performing an upgrade, you must disable vc-ws1a-broker or the upgrade will not work.
What a mess!!!