@dustinb3403 said in Verifying MS SQL Server 2017 Licensing:

@jaredbusch said in Verifying MS SQL Server 2017 Licensing:

@scottalanmiller said in Verifying MS SQL Server 2017 Licensing:

@jaredbusch said in Verifying MS SQL Server 2017 Licensing:

@scottalanmiller said in Verifying MS SQL Server 2017 Licensing:

@jaredbusch one of the complications is that there IS no virtual core. vCPU is NOT core.

a vCPU has vCores. Always. It might just be one. That is how it works.

Not in any system I've seen. What people call vCores are actually vCPUs. The vCPU might tell the OS it has multiple cores, but the idea of a vCore has never existed, only vCPUs. Vmware, KVM, etc. all the same. Core means physical, it's like having a physical virtual, it cancels itself out.

I am almost certain that VMWare lets you make a 1 CPU VM with 2 cores.

Hyper-V just says virtual processors.
0_1536703623096_9bebc766-93a2-498e-aa75-f621eb5bb0da-image.png

KVM says CPUs.
0_1536703642847_0252def6-e3e4-40f0-93ec-98417b2bb6ff-image.png

But I very clearly remember some hypervisor letting me specify a vCPU and vCores.

XenServer and XCP-ng also allow this.

1cpu 2 core etc.

Topology lets you state presented cores, not vCores. Totally different things.

@brrabill said in Nextcloud replacing OneDrive:

@scottalanmiller said in Nextcloud replacing OneDrive:

@brrabill said in Nextcloud replacing OneDrive:

Hmmm, sounds like one of those "MS messed up" scenarios we were discussing in the backup thread(s).

Yes, but it's also a "it didn't lose them their data" because they had local copies, scenarios we were also discussing. So a good example of both.

Actually it sounded like MS found them and put them back.

Local copies would also have been removed if they were "deleted" at the MS side.

Good point. The real problem here is the "customer shaming" that MS always does. Anytime MS is at fault, they try to shame the customer. Whether it is through official channels or "look the other way" as their engineers go into communities and do it. MS employees are out there attacking any customer that dares to mention that they've had a problem.

Like in this thread. Even once MS had admitted the mistake, and fixed it, an MS employee was still on the offensive trying to shame the customer and went ballistic once it was obvious he didn't even know how MS products worked.

@phlipelder said in Is Spamhaus the DDoS Arm of Microsoft:

@scottalanmiller BTDT

Back in the day the general sentiment was that RBL "services" were no more than extortion rackets. IMNSHO, that has not changed much.

With the advent of SPF, DMARC, and DKIM their relevance will become a lot smaller which is a happy.

SPF has been around forever, but a lot of mail domains still do not have it setup.
DMARC has barely any traction by comparison.
DKIM is not even a thing that does anything. DKIM only tells the receiving server what to do with a message that fails SPF or DMARC, instead of the receiving server deciding for itself.

@nerdydad said in 3 steps to maintain the 3-2-1 backup rule:

@scottalanmiller said in 3 steps to maintain the 3-2-1 backup rule:

@nerdydad said in 3 steps to maintain the 3-2-1 backup rule:

@oksana Is there a way of pointing Starwind VTL to a private cloud provider instead of Microsoft blob, Azure AWS, or Backblaze?

It's not that easy. What API is that private provider using?

¯＼(ツ)/¯

Then how would they point to it? It's like asking "can I use a thing"? I don't know, depends what that thing is.

@jaredbusch said in Microsoft buys GitHub for $7.5 billion:

@obsolesce The only reason I’m not moving my projects to githhub is because I cannot get private repository without paying for it. Gitlab offers that.

Yeah that's why I have everything on GitLab as well.

@scottalanmiller said in What Is Expected of Microsoft Server Support?:

Did they fix something broken with the OS that IT could not have fixed?

There are some quirky poorly documented things that are sometimes need to be done with ESEUTIL, and repairing from corruption, or failed log replays on Exchange and SQL. In theory it's not always Microsofts fault (caused by storage layer issues).

The other issue is driver issues for IO devices. In theory this should be transative (supported by teh OEM who agree's to support your OS). That said OEM's often don't support the free OS's because they don't have joint engineering commitements to fix issues on them.

RedHat has hard commitments from Avago/Broadcom to fix a driver HBA issue if they find it. Having a Server that has a supported OS means they have Avago's commitment to spend hundreds of thousands of engineering time to fix an issue. Redhat will possibly work to mitigate it from their side if possible. Missing the OS vendor in this conversation/effort can make things move slower or stall. Avago and other component manufactoreres will refuse support calls from a customer who bought the device through an OEM so without the OEM listing your OS as supported you are kinda screwed.

(A long time ago) I saw a case where a I/O device had outright buggy silicon, and a driver and OS side workaround was needed to resolve it. This is all the more fun as how the I/O device engineering teams work is they generally refuse to work with a customer directly and require the OS vendor escalate the issue to them.