Best posts made by 1337

@mroth911 said in Average Rate for Emergency Service:

I was just trying to help my family out. thats all. I know I am f'in crazy. But He just got the job and I felt that something like this they would like to fire him cause he was setup to fail.

That's very kind of you. It also puts things in a different light.

Does the customer know that you did any work on this? If that's the case I would just put the cards on the table. Give them a very good discount on your regular fee. Basically, aim low and get paid, instead of potentially getting nothing and perhaps also getting your brother in law fired. And the rest of the money you can in your mind write off as pro bono for your brother in law.

@Osvaldo said in Two or three screens. what is your choice? or just one.:

@Pete-S Hey Pete, 40" to 55" sound's crazy.

It's because most people are used to working on very small screens, like one tiny 24".

But it's not really as big as you might imagine:
https://core4.staticworld.net/images/article/2016/05/dell-4k-monitor-100662717-large.jpg
It's 43" Dell in the pic.

Monitors in this size goes for around $550 to $850 so they're not that expensive.

No need to wipe all the disks. It's a waste of time.

• If you wanted to erase the partition table you don't have to erase the whole disk. Just the first / last bytes of the drive. Or just remove it with fdisk or parted.
• If you wanted to remove superblocks after an earlier md raid installation you wouldn't need to wipe the disk either. You do: mdadm --zero-superblock /dev/sdX

You could also use wipefs -a /dev/sdX to remove just partition tables and raid signatures - if you have it installed.

And as soon as you put together a raid-10 it will start synchronizing the disk and write the same zeros one more time.

If you really, really wanted to wipe the drive so no residual data where on them, wipe just what is needed to make a raid-10 array out of it. Then wipe the array with zeros, as a block device. It will save a lot of time. As the synchronization process will write the zeros for you on the other drives.

@scottalanmiller said in GDPR Requiring Centralized Password Management:

@pete-s said in GDPR Requiring Centralized Password Management:

This is the GDPR. You can check yourself what it says. It's only 88 pages.
https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32016R0679&from=EN

Every countries in the European Union are required to make it national law.

Yeah, I've read most of it. But anything 88 pages is long enough to make creating FUD pretty easy to do.

Yeah, FUD is how the big boys make their money. If it's not fear, uncertainty and doubt then it's complexity. Make something that could have been simple, as complex and convoluted as possible so that you absolutely need lots of consultants and experts helping you. Which of course the supplier can offer. And finish of the cocktail of deception with a big chunk of vendor lock-in on top.

@guyinpv said in Synology one bad sector crashes whole volume RAID0:

I'm running them striped since I wanted more space and perhaps speed.

Disks fail, that is what they do. You're asking for it when running RAID0.

But now you just buy new drives and restore your backup.

WD has some test program that can verify that the disk is broken, then just send it in for warranty replacement - if it's still under warranty. WD Red had 3 years I believe but can be extended to 5 years for a small fee.

@matteo-nunziati I have a suggestion. You could potentially move your entire development environment off site to the cloud/colo/whatever. And then just use your computer to remote in and do the work.

With your uploads speeds it is much faster to just run remote sessions. And with that setup you are not dependent on your laptop as it is a single failure point for your business.

Or get a fiber connection or whatever it takes to get decent internet speed. 3 Mbps is not enough.

It's logical that linux is slower to boot since it supports more hardware in the kernel and has to do more checks to find out what it is running on. Could probably make it a lot faster by recompiling the kernel. There are embedded linux versions that boots extremely fast but only has what it needs in the kernel. And that goes for the rest of the services as well.

Windows is usually not 100% ready when it starts to show the login interface. So it appears to boot faster than it actually does.

But honestly though, I reboot my laptop maybe once a week if that. Doesn't sleep work in Fedora 28?

@guyinpv

When you work long enough in the same place, you start to think that you are responsible for it all. You know management don't have a clue. It's your systems that you manage. It's your solutions they are using. In reality, it's not.

It's their systems and they gave you money in exchange for your time. You are even with your employer each time you get paid. They invested money, you invested a piece of your life. Time you will never get back. That is the deal.

You have no responsibility for anything, except to do your job while you are getting paid. That is what being an employee is - a trade.

This is the problem you are having. You are confused about what the deal is and what your part is. Not who said what or what is reasonable, legal or whatever. But it's normal. It's why a lot of people suffer from occupational burnout. Now you just have realize it.

Are there any consensus on what VM hosts belong in the same pool when all hosts are running on local storage?

I'm using xenserver (xcp-ng).

I have two distinct usage for the hosts, one is for development and benchmarking and the other is for production.

There is no HA at the hypervisor level. VMs either have failover at the application level using load balancers/proxies or don't require HA at all.

@vhinzsanchez
Don't forget colocation as an option too. Usually a colocation datacenter has better internet connection, better power, better cooling and is physically more secure.

Ahh, @JaredBusch beat me to it I see.

We have a handful of workstations running windows but they are set up as "standalone" systems with local users, not connected to domain or anything, except a file server (running samba). Everything else in the network is linux/bsd.

What is the best way to administrate a small windows environment like this?

I'd like to do what makes sense but avoid Microsoft lock-in in general.

Your image is too small to see.

But also there is a difference between Mbps and MBps (more often written as MB/s).

If it's something like ASA5505 or 5510 you can do WAN failover but you need the right license. Security+ I think.

@WrCombs said in Windows 10 Reboots:

@Pete-S said in Windows 10 Reboots:

Could be hardware or software. Just replace the PC and install everything from scratch.

Or spend many hours on it then after that didn't work, reinstall everything and after that didn't work then replace the PC and install everything from scratch.

basically, but what do I know.
Im just the new guy.
im over it.

Honestly, if you're not the boss man, just do it the way he wants it done.

Assuming it's for servers you host, you are basically looking at routing at the level of a hosting company or an ISP.

Unless you own the IPs, meaning you have received them from IANA, you can't. If you own the IPs it involves setting up BGP routing and you need failover routers or it's all meaningless. That's about all I know because that level of networking is way over my head.

That's why colocation was invented. Put your servers in a datacenter and let the datacenter handle the redundant internet. And redundant power and redundant cooling and physical security.

Here is Firefox headless mode:
Firefox uses the same browser, just run with the headless switch. Works from version 57 and newer.
https://developer.mozilla.org/en-US/docs/Mozilla/Firefox/Headless_mode

I actually question the reasoning behind upgrading a R710 server in the first place. It's just too old. It has Nehalem 5500/5600 series CPUs. That's 6 generations old Xeons. Back when 4-core and 6-core where highest core count. Great if you like to spend money on electricity and cooling, otherwise not so great.

It would have been better value for money to get something newer and move the VMs instead. And low budget is not an excuse really as you can get refurbished servers with warranty for low prices. Something like a R720 for instance or newer.

So put new disks in a new/refurbished server and move the VMs. When everything is up and running and looking good, just wipe the disks and ebay the old server.

@DustinB3403 said in Having sluggish performance on my Xen Server VM's, looking for suggestions to boost performance:

@Pete-S That would have very easily been a much cleaner approach to this. I don't know if there was some sort of budget constraint but, clean is always the best if possible.

Agreed. My thoughts are:

• If you have no budget then do nothing
• If you have a small budget then make sure you get the most value for the little money you have
• If you don't have enough budget to get a good return on your investment then wait until you do

Some cost saving measures are just too expensive if you look at the total cost and what you get in return. More so for a non-profit that has to use each dollar wisely.

@gjacobse said in Spec'ing a new computer from Dell or?:

Just need a decent build

i5 or better CPU
16GB mem
500 GB SDD (could go with 250GB as the SSD I have now is 120)
Dual Monitor Support, HDMI or DVI
USB A and USB c would be nice,.. but not required.

That's run of the mill stuff. Just buy something and call it a day.

The setup with firewalls, etc looks like a lot like our colo setup except that we have more hosts but no end users and no POE switches.

Firewalls will need to sync states between each other which is usually on it's own interface. So an arrow between the firewalls.

If you provide services through internet you need some special sauce to have two ISPs use the same IPs. In our colo space they have BGP routing and redundant routers and redundant switches to provide us with that.

Your two switches should probably be stacked. Another arrow for that between the switches. Then you can run lcap active/active connections to your servers. You get twice the bandwidth that way. In a stack you manage all switches from the same place so it's easier too.

Don't you want to spread out the POE switches? Run two 10GbE links to every POE switch and place them close to the end users. Single mode fiber if there is any distance involved.