Linux Domain Controller
-
Tinc is a pretty neat mesh VPN which has less overhead.
You really want all static IPs for it though. -
@thecreativeone91 said:
Tinc is a pretty neat mesh VPN which has less overhead.
You really want all static IPs for it though.You always need that somewhere. Pertino handles it by actually being an elaborate, hosted hub and spoke system that mimics a full mesh. You can do the same thing with OpenVPN or even IPSec, just takes a lot of work.
-
@scottalanmiller said:
IPSec is definitely lower overhead when available.
IPSEC is lower overhead because it is has been offloaded in most cases.
If you are not offloading the encryption in a router, the overhead between the two is not all that different.
I use OpenVPN as the site-to-site method to connect the ERL at most clients. IPSEC always seems to have issues. For most SMB this is good enough as they will never saturate the OpenVPN link.
-
Yes, IPSec is such a pain to deal with. We used to have a huge site to site mesh using Netgear VPN hardware. That was well over a decade ago, though.
Now I feel old.
-
For the record, an Ubiquiti EdgeMax Lite will cap out an OpenVPN connection at about 10-14mbps because it will take up all the processor.
So unless you are going to push more than that, there is just no reason to worry about it.
-
Not talking much traffic at all.
Honestly might not even put all devices on it.
Going to still run local DC, so I'm thinking maybe three cloudatcost, my main desktop, and local DC all on VPN
So maybe 5, might push it to one or two more computers but nothing really intensive.
Would/Should I be able to host my dc on CentOS7 as well as OpenVPN? Or would that require two? (Dev1)
Thanks
-
Yes, you can put OpenVPN aggregator on the same VM as your DC, that's no problem.
-
This is more so what I meant.
Not typically in the 80's but it's usually floating around the 60's
(Just did a restart so I'm assuming thats why its spiking)
Edit: However; its definately not dropping from 87% now
-
Just doing a quick google search found this how to
I assume for the most part this should be fine minus not using google'd DNS records
-
@Sparkum said:
This is more so what I meant.
Not typically in the 80's but it's usually floating around the 60's
(Just did a restart so I'm assuming thats why its spiking)
Edit: However; its definately not dropping from 87% now
There is another thread on this. That is a completely worthless graph and should be removed from their site. It was created by someone who didn't know Linux and it uses the most common newbie Linux admin mistake that there is - reading the output of the "free" command incorrectly. Your system is not using 87%, likely more like 20-30%.
Show the completely output of **free -m"" and we will show you how to read the real memory utilization numbers.
-
@Sparkum said:
Just doing a quick google search found this how to
I assume for the most part this should be fine minus not using google'd DNS records
I would expect that to be fine. I run an ELK stack on Digital Ocean and our FreeBSD lab system.
-
Haha thanks thats alot better and brings me confidence in my box back.
Especially after everyone said a Samba server was one of the lowest resource wise.
-
@Sparkum said:
Haha thanks thats alot better and brings me confidence in my box back.
Especially after everyone said a Samba server was one of the lowest resource wise.
Argh, I hate that modified view. The simple "free" number is gone now.
BUT, the used number is correct. You are USING 65MB. Seriously, just 65MB. So you are in pretty good share. Out of 489MB total, 65MB is used. Leaving you with 424MB free!
Linux is awesome on resources
-
Gah!
Not sure what I did yesterday but now I cant contact any mirrors or ping anything!
Come on Linux be my friend!
-
@Sparkum said:
Gah!
Not sure what I did yesterday but now I cant contact any mirrors or ping anything!
Come on Linux be my friend!
nslookup or dig is going to be your friend here. Check your DNS settings. Something is not resolving.
-
Is there any chance this isnt on me?
When I first booted initially I got this as well.
And I dont beleve I've changed networks settings to date?
Just changed DNS (temporarily) to google's DNS and same results.
Also keep loosing connection to the console this morning.
-
Yep was definitely on their end, I wasnt able to get into the console for about 20 minutes. Was about to submit a ticket and there was a related article, which seemed exactly like my problem. Waiting a little more and everything works now (as origional settings)
-
I've never been able to full saturate gigabit links with OpenVPN. But I have with IPSEC and Tinc and that was using Core 2 duo era cpu's. Cisco stuff in $3k range couldn't do it.
-
@thecreativeone91 said:
Cisco stuff in $3k range couldn't do it.
That's like the lowest bar in the industry.
-
So how should I have my DNS settings,
If I leave them as CloudatCost default I cant ping anything, if I change them to what they would be once I get them added to the domain, nothing, if I change them to google's DNS I can see the outside world.
I assume this is not a good idea though? Or is there no real downside?
Would I have it set as google's DNS until I can get OpenVPN up and going?
