All storage on that server is internal so no drive bays. It's far from ideal. And the server doesn't have any hardware raid. Also since some of the ports are SATA-2 they are not suitable for SSDs because they limit the performance to 50% of max. So you are really limited to two SATA SSD drives.

That being said, storing video streams are one of the applications where HDDs are actually very suitable. Depends on how much storage you want but if you need more than a couple of hundred gigs, HDDs should be your choice.

Assuming you want lots of storage I would get two 3.5" drives and run md software raid 1 on those. Something like 2-6TB drives. NAS drives (like WD RED, Seagate Ironwulf) if you are on a low budget, otherwise SATA enterprise drives (Seagate Constellation ES or similar).

I would also get one Supermicro SATA DOM (tiny enterprise SSD that plugs into SATA port) and put the hypervisor and VM's OS on that. Just have a backup in case the DOM ever fails.
https://www.storagereview.com/supermicro_superdom_review

I wouldn't get SSDs for storage for your use case - unless your needs are very small or your budget is big. But if you have a large budget I would get another server with drive bays.

@scottalanmiller said in When is colocation the right choice?:

I'm in Europe so doesn't make sense to use US datacenters. But the quotes I have puts three 1U servers at the same price as 1/4 rack.

Wow, the difference here is usually huge. Like $50 - $100 for a 1U or 2U, and several hundred for quarter racks.

Yes, I guess the choice would depend a lot on the pricing.

I also have specific 2U servers in mind which would make it lower cost with a 1/4 rack already at two servers.

Each refurb server will have 128GB RAM on 20 cores, expandable to 256GB RAM on 40 cores. I haven't run benchmarks on it but looking at geekbench, machines with the same cpu have a score of around 3000 for a single core. So performance should be similar to most cloud providers that are not running latest gen servers. Hopefully the balance of physical cores and RAM will turn out fine, 6.4GB/pCPU. Of course I could allocate resources as I want and add another server if need be.

@aaronstuder said in Changing FreePBX Contact Header:

0_1537195564945_b6819afa-248f-4f7c-966f-69f8b3af8cc3-image.png

Looks like I am using Chan SIP?

That is not really a problem with anything.

We have a set of Group Policy Objects that are set up at the domain level with WMI filters in place for two of the three.
The first has all of the key settings for updates that are shared across the board.
The second using a WMI filter to point only to desktop operating systems with the necessary settings in place for all desktop OS versions that we manage.
The third is for server operating systems only. It delimits to download but not update automatically along with a few other settings.

We've been using this setup since I can remember and not once have we hit servers auto-installing and rebooting. It's too risky for that to happen.

@killmasta93 the directory had just been missing?

@black3dynamite said in Nextcloud 14 upgrade follow up warnings:

Is this necessary when doing a new install too?

Based on the message, I would assume no. But I have not installed a new 14 instance.

@jaredbusch said in Pi-hole on Fedora has issues with SELinux:

https://mangolassi.it/topic/17834/pi-hole-web-interface-not-working/15

That worked.

@jaredbusch said in Pi-Hole web interface not working:

On the other system I created a blank external.conf and it works now.

sudo touch /etc/lighttpd/external.conf

Got it

You can use Powershell as well

Set-MailboxAutoReplyConfiguration -Identity username -AutoReplyState disabled

@stacksofplates said in Fedora 28 Guacamole base install.:

@scottalanmiller said in Fedora 28 Guacamole base install.:

@stacksofplates said in Fedora 28 Guacamole base install.:

@scottalanmiller said in Fedora 28 Guacamole base install.:

@stacksofplates said in Fedora 28 Guacamole base install.:

@scottalanmiller said in Fedora 28 Guacamole base install.:

@stacksofplates said in Fedora 28 Guacamole base install.:

@scottalanmiller said in Fedora 28 Guacamole base install.:

@irj said in Fedora 28 Guacamole base install.:

@scottalanmiller said in Fedora 28 Guacamole base install.:

@irj said in Fedora 28 Guacamole base install.:

I see no value in Guacamole, personally. No Machine is a much better option.

Doesn't deal with the web interface. NX requires a client. Guacamole is as much a front end to NX as it is to any other protocol.

I wonder how often you really need people with no technical expertise to connect remotely to a linux machine?

For example, I have users on Chromebooks that need remote Linux access. \

SSH or RDP. I agree with @IRJ, it (Guacamole) just seems like a hassle.

RDP is pretty cumbersome in some cases for that stuff. And does that mean no NX, or having to maintain multiple for each box?

I definitely find Guac worlds easier. For one thing, it maintains a list of available resources. So the access listing is centralized, rather than making people maintain it on desktops.

I've never found d RDP cumbersome. I'm not talking about using NX on a Chromebook. Anything that isn't a Chromebook, NX is easier. For the Chromebook people SSH is native RDP is easy.

This whole thread is around how Guacamole isn't easy. You also don't need this to have a centralized list. If you're using CM you're going to have a list of systems anyway.

CM?

Configuration management.

Okay, wasn't sure. Yes, pushing out NX configs shouldn't be too bad, in theory.

NoMachine finds all of the systems in a network. For example, if you're on ZT it will list all available clients on that network. I was saying if you just wanted a list for people to SSH/RDP to, you already have it.

What we want, ideally, is not to use something like ZT (although it's an option that we will kick around), and not to make a list for everyone. But to have a portal (or something like a portal) that lists all of the machines available specifically to the person in question. We have specific pools and assigned boxes. For example, Maggie has a dedicated machine just for her, I don't want to see it in my list. And she doesn't want to see mine. But we both want to see remote machines dedicated for something we share.

@pete-s Those where taken using chrome but all browsers have the tools.

Here are couple screenshots from firefox webdev tools
0_1536977899506_Screenshot from 2018-09-14 21-15-09.png

Performance Tab
0_1536977990636_Screenshot from 2018-09-14 21-19-22.png

From the google chrome dev tools you can even get an audit using Lighthouse

You get a report like this:
0_1536978363389_Screenshot from 2018-09-14 21-23-46.png

@tx1tguy said in Cheap SSL vs Other SSL?:

Holy crap.. This is awesome. Thank you so much. Only thing that suck is the 90 day reissue, but who cares! Its free!

Cloudflare has multi year free certs if you're concerned about that.

@scottalanmiller said in Network error Windows 7:

@wrcombs said in Network error Windows 7:

@dashrender said in Network error Windows 7:

@wrcombs said in Network error Windows 7:

@scottalanmiller said in Network error Windows 7:

@wrcombs said in Network error Windows 7:

@scottalanmiller said in Network error Windows 7:

@wrcombs said in Network error Windows 7:

Anyone ever see this error before? do you know what it means ? how to get rid of it?
0_1536843443200_netowrk error.PNG

No, because I carefully never do that, lol.

I didnt do this, this is from someone saying " This is how I think it goes" and changing settings.

Also something to avoid 😉

Random people who don't know the basics, making completely crazy configuration changes... that's a disaster waiting to happen. Imagine if they were doing that on a PCI network or something 😉

They did! we use 2 NICs, one for internet, one for the PoS system so they are "seperated" the best they can be. They stacked and added IP's to the PoS NIC, So all of the PoS were accessing the internet...

Sure, but the internal POS network doesn't have a gateway to some place else inside it, right?
So you don't have a gateway on that network interface.

I dont believe there is a gateway, its just communicating within the PoS and the back office.

If that were the case then I shouldn't be getting this message correct?

Shouldn't. The warning seems like it would be genuine - it's warning you that someone messed up the configuration.

When I used to have these type of setups, I would set the gateway to the POS and set the metric to 10 for either that NIC or that subnet. That's assuming that you're talking directly to the one and only POS and nothing else on that network. While you don't have to alter the metric, it helped stop Windows from wigging out seeing two NICs or two IPs and two gateways on one NIC. If you experiment, you can always put it back. Different VASCs and different brands & processors specify what's ok with them. Because PCI.

@irj said in Edge (Sync) Favourites Toasted:

Edge has pretty much shown us that Microsoft will never be able to build a decent browser. In fact, they should just give up IMO.

I thought Edge showed that they HAD given up.

@networknerd said in VMware vSphere 6 Showing No CPU or Memory Used:

. I thought the back end data from the screenshot above was populated by data from the hypervisor only and not from VMware Tools, but we shall see.

I would have expected so, given the type of data that it is.

ZT is an SD-WAN, a specific type of highly complex VPN. Basically it is a hub and spoke VPN, but with advanced detection to allow full mesh ad hoc between nodes when appropriate. But at the end of the day, each link is a standard VPN connection. It's just how it is hosted (central controller) that is outside of the network of connected gear; and the hub and spoke vs full mesh auto detection that makes it so unique.

A nearly identical system is Cradlepoint, but it does not allow the mesh option, only the appearance of it, so that centralized traffic control on the controller is available.

Its a Xeon W, I heard. Which may be why you don't see it on the list.

We ran a bench mark against it recently, and it was pretty obvious that it was a Skylake just for the performance jump.

@dustinb3403 said in ResourceSpace - DAM:

Does anyone here use ResourceSpace

Just curious, nothing else.

Never heard of it, But was definitely worth looking into!

Thank you for the help. By posting the screenshots, I realized I had a typo in the file extension. .png vs .jpg
Once I double checked them all it now works.