@JaredBusch said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2014 on same host:

@JasGot said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2014 on same host:

Four 1TB SSD drives in RAID10 for SQL Data and VMs

IOPS is not likely your problem if your RAID card is configured correctly.

But

@JasGot said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2014 on same host:

Two 500GB SSD drives in RAID1 for the OS

This is likely splitting your RAID card's capabilities.

What card is it?

Edit: This is less of a problem with SSD because there is almost no read/write delay unlike with spinning disks.

H740P

@DustinB3403 said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2016 on same host:

@JasGot said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2016 on same host:

I've had chats with the developers, and they suggest the SQL server and application are are optimized. Leaving me with SQL Query issues in and out of the VMs.

The devs are idiots who work in just their silo. They know the software they are working on, and that's about it.

The host and its underlying hardware is where you are expected to be the expert and to pinpoint the issue(s).

I'm guessing you currently don't have any sort of system monitoring in place, besides what is a default on Windows Server 2016 Standard (no special tools). So you should consider setting up something like Zabbix to monitor this hypervisor, that way you'll be able to state with fact where the slowdown is, be it a IOPS issue (very likely) or a different contention issue with the RAM or CPU.

Zabbix requires the host be installed on something other than Windows. I'm going to try and compile statistics another way. I'll come back to Zabbix if I am unsuccessful.

@DustinB3403 said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2016 on same host:

What kind of IOPS are you getting from the hypervisor?
How is the server configured?
How many concurrent users, 10?
Have you monitored the performance to try and determine where the slow down is occurring?

No idea on the IOPs. I'll install "Zabbix" mentioned in your other post and report back.

Dell R540 Server with the following specs:
Two Xeon Gold 5117 2.27GHz 14 Core CPUs
(2 sockets, 28 cores, 56 logical processors)
128GB Ram
Two 500GB SSD drives in RAID1 for the OS
Four 1TB SSD drives in RAID10 for SQL Data and VMs

Windows Server 2016 Standard
MS SQL Server 2014 Standard

Server Runs AD, DHCP Server, DNS Server, SQL, and the VMs.

VMS:
Nine Windows 10 Enterprise VMs
PCs:
Six Windows 10 Pro Workstations also accessing the SQL Data

@scottalanmiller said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2016 on same host:

Check the network setup, make sure that the database server and the Windows 10 client VMs all have proper network drivers configured. Could be a point of issues if not.

I'll check this.

@scottalanmiller said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2016 on same host:

@JasGot said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2016 on same host:

I've had chats with the developers, and they suggest the SQL server and application are all optimized. Leaving me with SQL Query issues in and out of the VMs.

No way to know BUT the real issue is that it is slower than the same system before. That's the key here.

Can you describe the former setup?

I minced my words and created confusion. It's all new hardware and design. Previously it was SBS2011 with SQL installed and another server running Windows Server 2008 with Terminal Services. Both with spinning 10K drives and 32GB ram. It wasn't fast, but it was faster than this.

@Dashrender said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2016 on same host:

How much RAM do the VMs have assigned?

The Dom0 Windows instance can use all the RAM not assigned to a VM ( I think ), so that's likely one place it could be winning, depending upon how the RAM is divied up.

I had them set for 6GB Dynamic Ram They never use more than 2GB at peak, usually around 1.7 or 1.8. I changed a couple of them to 6GB NOT Dynamic and did not see any change in the user experience.

I also had them set for two vCPUs with 10% for VM Reserve, VM Limit, and Relative weight.

I tried 3 vCPUs, no change, and then I went to one vCPU with 100% VM Reserve, VM Limit, and Relative weight. I thought I noticed an improvement, but I can't be sure via the user interface, hopefully Zabbix will shed some light.....

@Dashrender said in Slow MS SQL Queries between Windows 10 VMs and MS SQL 2016 on same host:

How much RAM do the VMs have assigned?

The Dom0 Windows instance can use all the RAM not assigned to a VM ( I think ), so that's likely one place it could be winning, depending upon how the RAM is divied up.

I'll list what I know about the memory usage, not sure how it is spread about:
Total Physical RAM: 128GB
In use Compressed: 58GB
Available: 70GB
Committed: 60GB/147GB
Cached: 35.5GB
Paged pool: 592MB
Non Paged Pool: 1.92GB
Hardware Reserved 318MB
Speed 2666MHz

I haven't seem RAM usage above 50% and I have not seen CPU usage (taskmanager) above 7%

I'm going to go and grab Zabbix now.

I have a small office that has a single server. It has Windows Server 2016 STD installed (full GUI) and has Hyper V enabled.

MS SQL 2014 Standard in installed on the physical server along with some Windows 10 Ent VMs.

I have our application installed on server and also on the VMs. The SQL efforts of the VMs are slower than on the Server OS.

I've had chats with the developers, and they suggest the SQL server and application are all optimized. Leaving me with SQL Query issues in and out of the VMs.

I am looking for some advice on how to troubleshoot this and some suggestions on things I can do to try and get better speed for the end users.

This is a new and significantly more powerful setup than before, but is run slower than before

I need some help sorting out the exchange connectors on a server.

We were called in to figure out a spam problem. It's not severe and during our exam, we are trying to check the configuration of Exchange and get mail test tools to be happy.

Right now we are getting: Reverse DNS does not match SMTP Banner from mxtoolbox.

This is a simple MS Exchange setup; One Exchange server on a domain, people connecting with MS outlook from inside the building on the domain and from outside the building using Ms Outlook, and two copiers that can send e-mail.

The issue is all the receive connectors. It's not clear to me which one mxtoolbox is seeing.

They are:
Client Frontend EXCHANGE
Client Proxy EXCHANGE
Default EXCHANGE
Default Frontend EXCHANGE
Outbound Proxy Frontend EXCHANGE
Scan2EMail

I'd like to know what each one represents. ie; I presume Frontend mean Internet facing, but I'm not sure. Also, do internal MS Outlook users connect to a receive or send connector? (or both?)

Thanks.

@scottalanmiller Would you accept "Months"

My VM Snapshots are primarily to quickly rebuild NC after I mess it up!

@scottalanmiller I snapshot my VM occasionally. This is clearly not sufficient since I upload to NC hourly....

@JaredBusch said in NextCloud Disaster Recovery?:

@JasGot said in NextCloud Disaster Recovery?:

The one thing that still worries me is the fact that the files are not stored with human readable (or even original) filenames. A non recoverable NC install leaves me without any idea of which file is which. Unless I am missing something......

Are you using the Nextcloud encryption? If so, that is why you cannot see files.

By default everything is stored exactly as it is on the web interface / sync client.

Aha! Moment here..... It's because I use S3 as the primary storage...... Local primary storage is indeed native filenaming conventions. Thanks for opening that thought @JaredBusch

I have a NextCloud installed and running at a Linux server, and its configured to have Amazon S3 bucket as the primary storage and if I sync files at local folder its uploaded to the S3 bucket, but the problem is that the files appears as urn:oid:1158

Now, does this change any of your thoughts on just making sure the NC database is backed up?

Assuming I am forced to recover from the Cross Region replication, and no database backup; you can see how I am out of luck. So, with my primary datastore being S3, that database is even more important.

Just curious what you are doing to recover from a complete failure and breakdown of Nextcloud.

My NC is using S3 storage and is Cross Region Replicated with Revisions. As well as duplicated via a Desktop client with full sync.

The one thing that still worries me is the fact that the files are not stored with human readable (or even original) filenames. A non recoverable NC install leaves me without any idea of which file is which. Unless I am missing something......

What do you do to safeguard your data in NC?

@scottalanmiller said in Remote viewing software:

@JasGot said in Remote viewing software:

@wrx7m said in Remote viewing software:

@JasGot said in Remote viewing software:

We use ConnectWise Control (formerly ScreenConnect [SC]).

Honestly, I would be hard pressed to give any other option a second look. BUT, it didn't start out that way. In the product's infancy, and early in our learning curve, we only used its basic functions (which every other products offers too). But as we mastered it; and it became a more complex and flexible product, we have turned it into a be-all/end-all for RMM. The things we can do through SC would amaze you. Granted most of it is OS related, so it's not something SC magically provides, but the fact you can to it without ever seeing or connecting to the gui at the other end make it extremely valuable.

Are you using hosted or on-premises?

On premises. About 50 feet from me Running on an Intel NUC

Windows or Linux? Windows runs better with lots of resources, Linux runs way better for us in cloud on a "per dollar" basis.

Windows. Runs great. About 3500 Clients in Access Mode.

@wrx7m said in Remote viewing software:

@Donahue said in Remote viewing software:

@wrx7m said in Remote viewing software:

Screenconnect for Windows works well. We have it hosted by them.

You can also control mobile devices. My only complaint is that the Android app used to remote in to other devices, is not that great. The mouse control is the main issue. They recently updated it to be better than before, but it is still not great.

can you remote into phones? I constantly get calls from people out in the field that have problems with their iPhones, but have no way to help the remotely.

Yup -
https://www.connectwise.com/software/control/remote-support/compatibility

Not the iPhones, unless Apple changed their rules. You used to be able to remote control the iPhones, but they had to remove that a couple of years ago.

Android? No problem!

@wrx7m said in Remote viewing software:

@JasGot said in Remote viewing software:

We use ConnectWise Control (formerly ScreenConnect [SC]).

Honestly, I would be hard pressed to give any other option a second look. BUT, it didn't start out that way. In the product's infancy, and early in our learning curve, we only used its basic functions (which every other products offers too). But as we mastered it; and it became a more complex and flexible product, we have turned it into a be-all/end-all for RMM. The things we can do through SC would amaze you. Granted most of it is OS related, so it's not something SC magically provides, but the fact you can to it without ever seeing or connecting to the gui at the other end make it extremely valuable.

Are you using hosted or on-premises?

On premises. About 50 feet from me Running on an Intel NUC

We use ConnectWise Control (formerly ScreenConnect [SC]).

Honestly, I would be hard pressed to give any other option a second look. BUT, it didn't start out that way. In the product's infancy, and early in our learning curve, we only used its basic functions (which every other products offers too). But as we mastered it; and it became a more complex and flexible product, we have turned it into a be-all/end-all for RMM. The things we can do through SC would amaze you. Granted most of it is OS related, so it's not something SC magically provides, but the fact you can to it without ever seeing or connecting to the gui at the other end make it extremely valuable.

At the request of another user, I'm going to try this again. With a couple of pre-discussion statements.

We have been doing TS for decades. We have been VMimg Servers for years. We are now beginning to move TS customers to Windows 10 Pro hosted by Hyper-V.

This question has nothing to do with the design or deployment of a VDI solution.

When you deliver a physical demo of VDI to your customer or potential customer, who has absolutely no idea what the words Virtual, VDI, VM, Hyper-V, ESX, VMware, or any other word you are thinking of right now, mean..... What is your model?

What is your existing physical on site demo unit look like? What does it connect to?

The physical onsite demo for the customer cannot be a PC or anything that any normal end user would think is a pc.
It cannot be a laptop, or chrome book, or tablet, or ipad.
It must support 2 or more 23" LCD displays.
It must support a local printer, a local barcode reader, a local Driver License reader/scanner, a local utility meter reader, and an electronic signature pad.
The physical onsite demo for the customer must connect to local LAN resources for proof of concept.

In the final solution, there will NEVER be a cloud based solution for the customer, all connections and data must remain on site.

I don't give two hoots about your ideas on how to design or deploy a VM or VDI (call it whatever you like) environment. I only care about what you carry into a customer site to provide a demo of what will be at the users desk after their current PC goes in the dumpster.

You see, if you don't carry hardware into the customer site and set it up to show them, (because they have no idea what you are talking about), you may not have anything to add here...... and that's okay.

I am looking forward to comments from folks who provide this kind of a demo for their customers.

I'm on to other things now. This one seemingly simple and casual question has resulted in a rise in blood pressure and anxiety that I find completely unnecessary and unproductive.

I have found no helpful or beneficial responses so far here. Only argumentative replies suggesting I have already pigeon-holed my demo tactic and that I chose wrong.

@scottalanmiller The customer MUST see their local printer, Electronic Signature Device, Meter Reader Scanner, Barcode Scanner, and a sundry of other business specific peripherals, Software and workflows function correctly before they agree to anything. The demo need to function for a time of up to a few weeks.

I'm not looking to argue about ANY aspect of VDI, just curious how others are doing it.

@scottalanmiller You spend the vast majority of your days with IT professionals. I spend the vast majority of my days with people who aren't completely certain of the difference between a mouse and a keyboard, or SD and HD, or a copier and a printer.

If I had a nickle for every customer that "...knows what VDI looks like and how it acts, hard to find anyone that doesn't use it or something that looks identical regularly." I could buy a single piece of nickle candy.

Not in the real world: "Showing off a physical thin client hardware device is kind of pointless as you can describe it perfectly... it's exactly like a normal computer running the thin client software, but with a smaller computer that costs "less" (often they cost more.)"

@DustinB3403 Maybe we should get back to the base question, since I already know how VDI works. I'll rephrase the question. When you walk in the door to demo VDI, what do you have with you?

@scottalanmiller You need to type SLOWER, I was answering your earlier comment when you posted again, and again, and again....