Azure Local is the umbrella under which on-premises hybrid becomes a reality for many companies that bought-in to the Cloud First/Cloud Native mantra. The move back on-premises has been ongoing for a number of years now. Why? Billing for one. Unknown charges. Charges up. Charges down. No consistency. Exorbitant cost. Orders of magnitude more expensive. Lift and Shift. That's the most expensive way to become "Cloud Native". By moving all of those VMs back on-premises that shiny new Azure Local cluster is paid for in months. With a 60 month life span that means the company is making money hand over fist after the return. AVD = Expensive AVD + GPU = Super Expensive Cloud PC = Super, SUPER, Expensive Bring them back. Put them where they belong. On-premises. Save five, six, seven, or more every month. Give bonuses to IT! ;-) Oh, and that gut reaction to the opposite messaging to Cloud First? Yeah, that's the marketing triumph of the ages. Convincing folks they need something they really don't need to "modernize" their business. But, you're more "Agile" in the Cloud! ;-)

We've done a few cost comparison run throughs with companies that have most of their infrastructure in the public cloud. It's orders of magnitude more expensive to run those workloads in the public cloud. Time Slicing, that is breaking those workloads down to apps, is still more expensive than running the whole backend and frontend in VMs on-premises. Public Cloud is _always_ going to be a lot more expensive.

Office LTSC Professional Plus 2024: No more Publisher While it's really not an important thing for most of our clients there are a few that use it for desktop publishing. So, what's the recommendation for an alternate?

@sysadafterdark On-premises is the only way to guarantee my data stays my data and is not "anonymized" to be sold to the highest bidder.

I think Dell was one of the first that mastered the modular DIY attitude to relieve stress on front line support by having the _user_ run the first troubleshooting steps. Necessity drives innovation. That being said, it looks like 45 Drives has adopted those excellent lessons.

@theownerop The problem: Auctions show a lack of interest to put in the time to learn the business in the first place. It's truly sad how much knowledge has been lost over the last 20-30 years especially. Folks built the SR-71 & Rocketdyne F-1 with their hands, brains, & a slide rule. :-(

TROUBLESHOOTING: TASK MANAGER - PROCESSES TAB The pic shows the columns we have set up in Task Manager by default. In this case, we're looking to see why a QuickBooks backup is taking so long and why the file is so large. What does the image tell us? The VM has 4x vCPUs and 16GB vRAM assigned and is on a Storage Spaces simple space which is equivalent to a RAID 0 stripe setup. Okay, so we've discovered that QuickBooks is running single threaded since it's hitting 25% CPU. We've also discovered that the Disk I/O is pretty significant at times. Since the folders within the company file location were oddly named it was assumed (yeah, I know) that the backup would be quick and tiny. It turns out that there's a secondary folder set with product images in it that seemingly had nothing to do with the company file. 12GB backup file later we're in business! :0) If we still hit roadblocks on figuring out where the apparent stall was coming from the next steps would involve ProcessExplorer and/or ProcessMonitor.

@willdepue Heh Ask an old timer about flame wars back in the BBS days. Nothing has changed. Anonymity is like a couple drinks: The social inhibitions go out the window.

That being said, what are the disks? SATA? SAS? NVMe? Hybrid? What's their makeup? Is there an HBA or RAID? Make/Model? How many 10GbE ports on each node? Is there a switch or are there switches? If yes, do they have enough bandwidth to cover a busy cluster? Jumbo Frames enabled? If yes, has every port end to end been checked to make sure MTU is set correctly? (This is one of the higher up reasons for storage performance issues)

@qwertymodo It's on my To Do List to run S2D on a 4-node setup we have on our build bench to get a good picture of its performane metrics. Then, to run Proxmox/Ceph and run as close to similar tests as possible as I've heard performance on this platform is not the best.

@SwiftOnSecurity When did Utilman.EXE become a thing?

@zk54 Heh ... making the world a better place one stroke at a time! Awesome stuff! :0)

@UK_Daniel_Card I haven't looked into what's out there for Remote Desktop attacks in a long time so TSGrinder is probably "Olde Schoole". ;-) Thus, it hadn't dawned on me that the attacker's tool chest could be augmented. Thanks for pointing that out. :0)

@UK_Daniel_Card TSGrinder has been around a long time and as far as I know can still grind away with Network Level Authentication (NLA) enabled. Better to have RD Gateway set up with a 2-factor authentication setup. We use DUO.

FREEBIE: URL REVEALER I despise tiny URLs just as much as I despise QR Codes. No, I don't want your tracking beacons following me around tracking my every move thank you. And no, I will _NOT_ scan your QR Code for anything I don't care what it is.

@SwiftOnSecurity Heh ... I LOVE it! Who can take that seriously right? IYKYK :0)

@d0rkph0enix I'm glad to hear it! :0)

@liveindetail Automatic (Delayed) does not rely on anything else. It just does what it's supposed to. Restarting the service does not always work as there are dependent services that also need to reboot. It's preferable to get it right from the start IMNSHO.

@qwertymodo What's the storage fabric and unit setup?

@keviniddles We started with DCs back in the day when it became apparent that we were being locked out by something ridiculously simple. We started setting it on all Windows servers deployed after maybe a half dozen odd times where NLA did it on a non-DC. Set it. Forget it.