Please share 
@jfranco
Maybe my joy was too early…
I think you know this… Unfortunately, PCNS cannot be installed on VMware. javaAPI is disabled in free VMWare, so PCNS cannot stop free VMware.
But PCNS is able to run a script as a result of the events received from NMC, although this is not supported by Schneider Electric lately. There is no script in the PCNS source and it is not documented.
The solution was that I wrote a script that can be used to remotely stop VMware and set PCNS to run when the event occurs. It worked great during on tests, but there was a problem because there was a power outage early yesterday and VMware did not stop. It turned out that I restarted the PCNS VM earlier and the script setting disappeared from pcnsconfig.ini.
I have since tested it and but the setting is lost when the PCNS VM reboots. It is included in the PCNS settings but disappears from pcnsconfig.ini. I don’t understand the reason, but I think that PCNS does not support running the script so that it is not possible to stop VMware by bypassing the javaAPI.
At the same time, I am testing the available versions of PCNS and the configuration options to find a definitive solution.
I don’t know if VMware and Schneider will close this backdoor if someone publishes a solution for this. Maybe it has already been closed and that’s why it doesn’t work reliably?
If I manage to solve this problem, I will apply.
Regards
Hi @jfranco,
How can I send you a private message? Could you send it to me so I can answer it?
I have a little problem with PCNS and VMware, maybe you could help me test it…
Thanks
Hi Steve,
Have you thought about changing to Xen, Libvirt or Proxmox?
Besides these related issues, you’ll also have problems with backups!
Thanks,
Hi,
I used Xen many years ago, later switched to libvirt. Now I use VMware. Proxmox has limited support for hardware RAID controllers, so I chose VMware. But nowhere was it described or warned about what to expect in addition to the restrictions on their website.
Now it’s too late to switch to Proxmox, it would be a terrible job that I can’t solve now. I can save the important stuff from the VMs and take a manual snapshot from time to time…
Proxmox CAN run your whole VMWare ESXi 1:1 as “Nested Virtualization”. I have a client using this for two Novell NetWare Servers, needed for Archival / Legal reasons.
This works rock solid, and would allow you a step by step migration for your VMs to Proxmox (More efficient than nested virtualization!), and you get the advantages of Proxmox Backups of VMs.
Migration of the VMs sounds complicated, but actually, Proxmox can read and use VMDK files, so it’s mostly a case of removing the VMtools, setting um the MS MergeIDE utility and converting the file from the command line… Depending on VM disk size, and your hardware, an hour is not a bad estimate (Per VM!).
Proxmox can use hardware RAID controllers well, but works better with ZFS. I have several clients still using hardware RAID on Dell & HP servers…
To err is human.
To learn from them is a sign of wisdom.
Making the same mistake twice is a sign of stupidity
Doing it three times is confirmation thereof!
My 2 cents
Andy
1 Like
@Andy_Wismer
This may be off topic, I apologize in advance.
I borrowed two used 10TB disks from a friend and I have a used 500GB SSD. I installed Proxmox 7.3 on these.
I’ve never used Proxmox before, but I’m trying to move VMs from VMware to Proxmox and I’m just hoping I’ll have enough disk space. If it succeeds, the VMs will run on it for the time being until I install Proxmox on the Dell server instead of VMware and migrate the VMs.
But I have a problem. I only see free space on the 500 GB disk and I don’t see where the LVM created from the two 10 TB disks is mounted… All I see on the Proxmox interface is that there is no free space on the vg0 LVM.
I can’t find where on Proxmox to set the uninterruptible power supply for shutdowns due to power outages. It doesn’t have NUT, does it have to be installed separately?
Is it possible to configure NUT on Proxmox like Nethserver to communicate directly with APC’s uninterruptible network card?
Thanks
Hi @steve
As you’re the original poster, it’s your decision how far off-topic you want to go…
Proxmox / Disks
Generally, it’s best to install Proxmox ONLY on the intended system disk(s), no matter if using real HW RAID or just simple disks or plan to use ZFS from boot on.
Proxmox will reserve 10% of that disk as LVM, the rest wuld be in LVM-Thin.
Proxmox uses XFS as standard (like NethServer) if you do not specify something else, like ZFS or CEPH!
LVM is for ISO Omages, Backups, ContainerTemplates.
LVM-Thin is for VMs and Containers.
Thin only allocates space, if actually used - this makes for faster backups, and less space used on the whole system.
Any other Disks, it’s best to format / install when Proxmox is up & running. You have the option of using ZFS or not, mirroring, whatever.
Proxmox / Nut:
Proxmox works very well with NUT, as server, client or standalone, it just does not have a GUI for this.
Proxmox can shut down all VMs correctly when being shut down via UPS. If no qemu-guest-agent is installed, this needs to use the BIOS of the VMs virtual board, not very reliable. Very rock solid and fast if the qemu-guest-agent is installed.
As Proxmox id fully based on Debian (Indeed, they’re one of the biggest contributers, especially in ZFS), installing Debian components is easy… instead of yum install it’s apt install…
Here a some of my personal notes to this:
————————————————————
NUT in Proxmox (Als Client)
————————————————————
apt-get install nut-client nut-server
systemctl status nut-monitor.service
systemctl enable nut-monitor.service
systemctl restart nut-monitor.service
nano /etc/nut/nut.conf
MODE=netclient
nano /etc/nut/upsmon.conf
MONITOR ups@192.168.151.80 1 monuser secret slave
systemctl enable nut-monitor.service
Check server with this:
upsc ups@192.168.151.80
————————————————————
————————————————————
NUT in Proxmox (Als Server)
————————————————————
apt-get install nut-client nut-server usbutils
systemctl status nut-monitor.service
systemctl enable nut-monitor.service
systemctl restart nut-server.service
nano /etc/nut/nut.conf
MODE=netserver
nano /etc/nut/upsmon.conf
MONITOR ups@192.168.175.62 1 monuser secret slave
systemctl enable nut-monitor.service
upsc ups@localhost
————————————————————
If you have further questions not yet answered here, drop me a PM!
As you see, I’m almost always around to help people on the right road / track.
My 2 cents
Andy
BTW:
Proxmox released version 8 a few weeks ago, it can be upgraded any time yyou want!
2 Likes
Thank you @Andy_Wismer.
How will I be able to send a private message if I have questions about this?
Click my Symbol / Image, then Message…
Left top of my post…
1 Like
I finally gave up on esxi 7 a few months ago as I could not get usb audio to work from a vm anymore. Passthrough worked in 6.5 but after days of googling I came to the conclusion that it has been removed by vmware for some reason in 7 and higher.
Now running proxmox and couldn’t be happier. Way more powerful with more features. Exporting esxi vms was surprisingly easy and my Nethserver vm was down for a very short time during the transition. Running mirrored zfs on all drives.
1 Like
Thank you @happnatious1 for your opinion and sharing your experience. I am now familiar with Proxmox, for now I am working on the APC UPS snmp-ups solution to make it work.
@happnatious1 personal and possibly flawed guess: VmWare wants to cut-off the uncertified realm of hardware that had been connected to ESX since the market deployment.
Starting with external, proceiding with internal (in fact i cannot add NVMe driver to ESX8, but i can on ESX7… so I installed on a Apple MacPro trashcan).
A hardware-partner request? Well… ESX 6.7 has been released to market in 2018, ESX 7 in 2020, 8 in 2022. VmWare inc. changed owner/mainshareholders few times, but still Dell Inc has some… stronghand in there. So having a close relationship with a enterprise-class hardware brand might… steer some decision to “better sell than support”.
Some people can say… “proxmox is an alternative”. I cannot agree.
Proxmox PVE is gaining a lot of traction and can do so many things close to ESX, but it’s not.
Takes advantage of two solid products as KVM and LXC, allowing to run both VMs and Containers, it appears to be really solid, capable. But as products that relies on PVE? I mean…
If you look into ESX backup products, I assume than at least 20-30 could be found, from GPL products to licensed software like Veeam. For PVE, outside integrated OS backup solution (like NethServer) excluding PBS … 5 names? Ok, underlying there’s debian, KVM, LXC, but the integration with Proxmox is not the same.
On the other hand, on ESX hardware device producers would like to… “cash” the ESX connection, like APC did. Unfortunately, VMWare cut too many ropes that were available in the past, so PVE might be a nice alternative, but don’t forget that under the hood, there’s Debian with an Ubuntu Kernel.
Considering what happened to RedHat few months ago… I’d love not to see Canonical play similar game for increase revenues faster.
Starting from ProxMox customers.
1 Like
Proxmox beats ESXi any time, flat!
VMWare has such a crappy supported hardware list, and has gone from bad to worse.
Most “Features” are paid for extras - yet still not as good as on Proxmox, which provides them all for free!
Where Proxmox easily beats VMWare:
- Backups
- Clustering
- Fast Migration
… and a lot more!
Best of all:
A simple licensing option, worth it’s price!
Almost all my clients use a paid subscription for Proxmox, at 100€/year per CPU socket it’s cheap compared to VMWare ESXi, Microsoft Hyper-V, or Citrix Xen…
How much do your clients pay for VMWare ESXi licensing? Or is it too expensive, you have to use the extremly limited free option?
My 2 cents
Andy
Is quite notorious (and refreshed a lot o times) that you are a ginormous fan of PVE bold phrase was a easy bet.
But “crappy supported hardware list” is a bit misleading as definition.
I can agree that is more and more restricted during the years, but this also helps reduce hiccups and headaches if you’re trying to push the limits of the platform. Still I don’t love the IBM/Apple way to manage hardware-software relationship (my software runs only on my hardware) but the most of the results about reliability and compatibility issues are definetively better than the average experience.
Unfortunately the relationship between VmWare and hardware vendor leads to pricey hardware to make things work.
On the other hand (paid to free feature ratio) I’d love to remember that as many others projects, they’re are standing on the shoulder of giants. Debian, LXC, KVM. VmWare might have took advantage during the history of the product on reading some sources, but had to pay a lot of coding work. And as far as i know, no one considers “cheap” software licenses, unless was paying a lot more before (Mainframe to OS400, for instance).
I’m not in the position to share customers licences fees. I bet that your customer, befoure paying proxmox, were paying ESX in several flavours. I hope that canonical won’t go whimsical about its kernel.
And we all also know you as an advocate for Microsoft - yet strange enough, you’re not using a Microsoft firewall…
My 2 cents
Andy
3 Likes
Advocate? Never see that coming.
What I always tried to say is “Microsoft sucks a lot less nowadays” (and you agreed), not willing for debate about the fairness of licenses and fees. I mean… last Patch Tuesday awarded the whole community of a massive laugh about “you should only install english OS and software” because Exchange patches won’t correctly install on different languages… but no project can be defined “bugproof” and “patchsafe”. Log4j still there to tell “you’re doing wrong”…
Two indicators:
NethServer 7.x, but also Proxmox PBS can update for years - without needing a reboot.
A well maintained Windows System will need at least a reboot once a Month, for the Patchday release of .Net, if not for the OS…
→ So higher avaiability and security for Linux / UN*X systems with less downtime.
And any halfway good hacking group will hack into the best protected Windows Systems like butter, see the US Government vs China (or Russia) here…
And as seen at a friend / client: Opening up a browser page on a home system (No Internet involved, the home system was freshly installed) can still cause a Windows System to get a Blue Screen!
Using Edge Browser AND Firefox with latest upgrades…
And the page worked on other PCs and Mac / Linux…
Trust Microsoft?
Trust is good, double-checking is even better!
→ Atributed to the good old Josef, aka (also known as) Josef Stalin!
My 2 cents
Andy
But more important: What about Winodws Firewall! Was the last one with SBS 2012?
In any case, this is a support Post, not a chat post, and from that point of view, we’re already very off topic now!
We could open a chat about this, but I really think that besides a few Die-hards who’ve never tried Proxmox, most here use Proxmox as Hypervisor, after, like me, having used VMWare for years.
And, the user in question here seems to just have moved from VMWare to Proxmox.
1 Like
…except that Neth, PBS, PVE, and anything else running Linux really should reboot when a new kernel is installed. e-smith/SME enforced that far too frequently; Neth and the rest do it too little.
A parallel issue comes up with TrueNAS a lot, where the recommended hardware list is also pretty narrow, though it will “work” (for certain values of “work”) on just about any x64 hardware with enough RAM. So which is better? To only work on a limited set of known-solid hardware; or to “work” on just about anything, including garbage hardware that’s flaky on a good day? My Intel X710 NIC was supposed to work under Proxmox–but didn’t; I had to downgrade to an X520. But sure, you can use your POS Realtek NIC. Particularly in a business environment, I think there’s a strong argument to be made that the VMWare model could be better.
Not to mention that PCIe passthrough is very immature in Proxmox. For lots of applications, that matters–but for some (like virtualized TrueNAS), it’s critical.
As you know, I like and use Proxmox–and I’ve got a bit of money in proper server-grade hardware to run it on (and I’d like to grab another pair of R630s when budget allows). But I’m pretty sure there’s a reason that ESXi is the industry standard beyond simple inertia.
AFAIK, NethServer excepted, but Proxmox PBS (Proxmox running LXC can’t pull this trick for obvious reasons, the LXC will crash!) does this fairly well according to best practices…
After installing a newer Kernel, PBS will reload (“boot”) from the newer Kernel using a sort of virtual initrd, until the newer Kernel is completly running of the newer system. This has been practiced for several years on AIX, for a long time (20y?) the ONLY UN*X able to update the Kernel without rebooting (and probably also reverse desgned for Linux a few years back) …
In my opinion, this is the best way to go, a decent compromise between reboots, availability and security!
You can also enforce a reboot for every editor update, like Microsoft does when Word (Office) is updated. Libreoffice doesn’t NEED a reboot to update a office packet! Yet it’s still easy to break into the latest and best protected Microsoft resources, as mentioned somewhere above…
Probably, by ijust installing MS Office, you’re inducing a massive insecurity in the system, any following reboot is just a placebo for peace of mind.
I dare say the MS systems used by the US government are among the best protected Microsoft systems out there (OK, maybe not every branch!)… Yet, apparently, the chinese can read that stuff “on demand”, like before a high level visit by Blinken or whoever…
Every farmer knows that monocultures aren’t good for the land…
Continuing the 60ies philosophy in IT: “No manager ever got fired for ordering IBM…”
And then came the two guys in a garage…
My 2 cents
Andy