@StorageNinja said in vSAS - Value Serial Attached SCSI:

100% of the time a SAS drive needs a SAS HBA or RAID Controller

Because the RAID controller is actually a tiny computer with its own built on SAS HBA 🙂

@scottalanmiller said in First Look Ubiquiti Unifi UXG Pro:

The new, unreleased Unifi UXG Pro just arrived here at the NTG Dallas offices. Woot! It's dual power supply, dual WAN, dual LAN, touch screen LCD and up and running!

More details as we get to play with it.

DAMN IT!!!

@Pete-S said in How Many HCI Nodes for the SMB:

But if you run a generic benchmark that is not designed to give inflated numbers, the situation is different.

I don't run generic benchmarks in production for a living thankfully 🙂
The reality is most CPU intensive stuff takes advantage of at least some of the new offload extensions and libraries. Also memory throughput is often the limiting factor for databases and other intensive IO applications.

@DustinB3403 said in Powershell variable help:

@Pete-S That is going to be complete subjective based on the type of spam that might come in.

I follow what you're saying but I'm kind of stuck between having something functional or not (due to this).

Well, you need to decide what you are allowing and not. For instance you have already decided to just use the subject line and not any of the other parts of the email, for instance author.

You need to run the string through regex or something like that to remove characters you are not allowing. It might take more than one string operation to sanitize. I think you can escape any special character you want to allow with backtick `.

Another option would be to allow the user to enter the entire KQL query themselves.

Also if you get an error when trying to apply the search you put out an error "Invalid search input!" and go back to prompt the user for input again. That would be one way so solve problems.

@BrownLuther007 said in Any good self-hosted discord alternatives?:

Hi, I might be a little late but specifically for the self-hosted, telegram is there. And for basic functionalities, there are Element, TeamSpeak 3, Overtone, Steam Chat, & Tox.

Telegram isn't self hosted, though.

@JasGot said in MSDTC issues:

Well, we found how to fix the symptom, be we do not know why it is so.

This network is 10.1.10.0/23

The only users having trouble were on the 10.1.11.0 portion of the network.

There are no configuration mishaps anywhere in the workstation, servers, switches or routers.

It's just weird.

But the MSDTC errors have stopped, and the staff who were suffering are happy.

I'm in a network transition myself, and I've run along a sorta similar issue.. definitely a PITA to troubleshoot and solve...

@scottalanmiller said in Proxmox install for use with a ceph cluster:

Yes, but it blocks SMART so you never want to do it, it undermines the stability of the JBOD. There's always a standard controller on the MOBO for the JBOD connections.

Also blocks TRIM commands (not that I trust the Linux TRIM driver to ATA drives given how many one off exceptions to disable they've had to write).
Operationally it's messy because on a drive failure you have to go in with PERCLI etc, and rebuild the RAID 0s. We used to run this but it was just a royal pain in the ass.

@MrCRuss said in Drobo B800i startup issues:

Had this just happen now. Where you able to save the data by moving the drives over to a new unit?

Another Drobo unit, yes.

Upgrading servers today.

But, like I said, if you are moving everything, you can simply backup and restore, and then update DNS.

@JaredBusch said in linphone: remove/hide “default identity”:

@scottalanmiller said in linphone: remove/hide “default identity”:

@JaredBusch said in linphone: remove/hide “default identity”:

@Dashrender said in linphone: remove/hide “default identity”:

@scottalanmiller said in linphone: remove/hide “default identity”:

@Dashrender said in linphone: remove/hide “default identity”:

@scottalanmiller said in linphone: remove/hide “default identity”:

Most enterprise PBXs, and FreePBX is no exception, give you the ENTIRE system. It's an "appliance." You don't download it as software and install on top of an OS as if it were an office suite or note taking application (although that's possible.)

At least not any more, and not for several years... but there was a time, not THAT long ago that you did.

Pretty long ago. Like mid-2000s I'd say. That's more than a generation in IT terms.

That doesn't seem right - I recall building my first FreePBX and that was only like 5-7 years max and you had to install from scripts - they didn't have a DL ISO for install.

If I had to guess, you did PBX in a Flash from Nerdvittles.

That was a scripted install on top of CentOS. But it was still nothing more manual than a single script.

I remember that. "In a Flash", haha.

Compared to the manual processes that existed before then, it was good.

In the end I had issues with the crap that the system pre-setup. It was all at the novice or hobbyist. Not business.

Definitely, it was always very hokey. Way, way too many gizmos and whatevers and way too little "feels enterprise stable."

@EddieJennings said in Script for Creating VMs from Template VM in KVM:

@travisdh1 said in Script for Creating VMs from Template VM in KVM:

@EddieJennings said in Script for Creating VMs from Template VM in KVM:

@Pete-S said in Script for Creating VMs from Template VM in KVM:

Not the exactly the same thing but you might want to look into how to create a VM from scratch.
Meaning a script that will set up a VM with vCPU, memory, storage, network etc and then boot it from iso and have it do an unattended install, create what users you want and install the packages you need.

That's one of the next things I'm looking into.

@EddieJennings Also remember about things like kickstart in RedHat based operating systems. In Fedora/CentOS/RHOS you can use a kickstart file to automatically select all the install time options for the OS. A short time later you've got a fresh server and all the time it took you to setup was running the creation script on your hypervisor.

One of the things I'll need to figure out going the Kickstart route is setting the hostname what I want it to be at the time of installation. Likely not difficult to do, I just have to figure it out. Or perhaps, I can just truly take the approach of just making a clean minimal install, and then later configure to whatever specific thing I'm wanting the VM to do for my lab / testing.

Inside the kickstart file you'll find something like this:

network --hostname=centos8-4.example.com

We use debian as our goto and then it's called a preseed file. The only real thing that can be tricky is to tell the installation what kickstart/preseed file you want to use. You can do it in different ways. If you don't want to rely on dhcp/tftp/pxe etc you can roll your own iso file. I think the kickstart file can also be mounted as a drive that the installation will detect when it starts.

I think the best approach is to make an automated installation with same basic settings and some of those will get changed later in the installation. For example you can use a fixed hostname that is later changed from ansible.

@brandon220 said in Browsing shares from W10 (2004) on a domain:

Thoughts?

Create a shortcut on the desktop for our user to get to the shares they need. If AD is setup, use GP to add them.

@Pete-S said in 4G Failover questions:

@travisdh1 said in 4G Failover questions:

@Dashrender said in 4G Failover questions:

@travisdh1 said in 4G Failover questions:

@scottalanmiller said in 4G Failover questions:

@Dashrender said in 4G Failover questions:

holy crap - it cost you $250 a month to run it at home - WTH?

How can it not? Ever priced out the power costs on a server? It's not cheap.

Yeah, power alone can make moving a server to a colo make a lot of sense.

Sadly, every place that offers 1u of colo space I've looked at has a max power draw of 120w and my current home lab box normally draws around 130w.

Time for a more efficient server?

Maybe in another couple of years. It's a Dell PowerEdge R620, 2 Xeon E5-2660 8core 16 thread CPU, 24 4GB 1333MHz RAM modules, 4 Cruicial MX500 500GB drives, Broadcom 4port BCM5720 Gigabit ethernet adapter, and a single 750 watt power supply. It's currently only drawing 112 Watts, so who knows, maybe I could "get away" with it.

120W is a very low limit, just 1 Amp. Colocation America has 2 Amps , so 240W, which low but still more reasonable.

BTW, your server config is not the most power efficient so there are things you could change if you wanted to that would drop the power requirement a lot.

In no particular order:

drop from two to one CPU replace CPU with E5-2600 V2 series. They use 22nm tech and have about 25% lower energy consumption in the same socket. Up to 12 cores. use low voltage CPU models, for instance E5-2650L V2 (10 cores) replace memory so you use fewer modules with higher density, for instance 16GB or 32GB use low voltage memory modules DDR3L replace the power supply with the titanium model

@notverypunny said in Automation thought:

Syncthing on the reports directory?

Definitely an option.

Okay - Can we stop now? Going back and forth isn’t healthy.

Why apt-get over apt

Okay, the question stemmed from a Raspberry Pi 3b+ running DietPi. I’m not going to elude that I know how much is or isn’t available in this system, but it’s stripped pretty well from the bits I have read.

It’s what the DietPi Updater used. It isn’t my typing the command ad hoc. So - you can quit the beatings as they aren’t going to help me any.

On my NC instance running 18.04 - I use sudo apt update. So,.. there is that.

apt install man
Could I install this? Yep - you guessed it, - I can. Do I want to, I’m not certain. Will I, again, I’m not certain.

Now - here is a cup of coffee for both you. This topic is closed.

Those aren't even business class companies. Focus on that... consumer service and total lack of support sold at crazy high cost to gullible customers. Verizon, Comcast, Cox, Intermedia... these are the anti-shortlist of companies. The ones no one evaluating anything would talk to. Anyone who knows the vendors and basic business approaches would have ruled these out without having ever talked to them. There are loads of good vendors out there, these are famous as the "never, ever talk to them scammers." Other than Intermedia, famous for having bad support and insane prices, the others are ISPs, which violates basic business common sense.

That's not a rant, that's explaining what your tactics should be based on. Sure they are big, that's not their problem. These are shitting companies preying on the gullible. They have no incentive to make a good product or support the customer. They aren't cost effective, they lack modern (meaning post-2000) protections and flexibility that's just assumed to exist in any real business, they don't have good support, they put the customer at risk.

You clearly stated fail2ban actions not iptables actions. They are not the same thing.

@scottalanmiller said in Fanvil possible firmware issue, non-standard port:

We've seen that kill fail2ban so that it ties up so many CPU cycles that performance drops.

That said, managing iptbales would be the admin's job. Monitoring the bans and jsut blocking entire CIDR would be a normal need.

Preemptive IP blacklisting is also a normal, intelligent thing to do. By geo, common known CIDR, etc.

There is zero reason to leave any PBX system, for a typical American SMB, open to the entire planet by default.