Difficulties with *.site.com and 443 redirects. How to fix?

Question

Nginx

Gentlemen, please tell me about nginx.

the config is formatted like this: https://pastebin.com/uWvg0LZv

and there is a difficulty: it doesn’t want to redirect to 443 and
when
test.site.by is requested, it
gives
https://%2A.site.by/
which, when decoding, I get
https: //*.site.by/

I don't understand where the plug is

Answer 1

my God and simply to register a host not destiny?
return 301 https://$host$request_uri ;
that's why you invent a bicycle
3 human configs for each domain separate beautiful convenient fashionable youth
No need to invent a bicycle, and then heroically solve it.
In short, do this
proxy_pass https://127.0.0.1:4443 ;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header HTTPS YES;
And most importantly, in 99% of manas on the Internet there is an error and you make it like idiots, then you get
problems with refery
http ->http
https -> https
without any crutches in the form of 443->80

Answer 2

If debian then proxmox.
I don’t know how it is there in the cluster, but how simple the virtualization management system works just fine.
I worked with about 20 proxmox installations, there are no special complaints. I myself keep several of my servers with proxmox and also just rejoice (for a year and a half, there didn’t seem to be a single problem from proxmox). The guys I knew raised a cluster with live migration on the third version and said that everything was already in production there and it works for them so far, which is very good.

Answer 3

If you already have MS and Vmware licenses, then choose between them.
There was a real experience of remote implementation to Krasnodar (full transfer from XenServer to Vmware with VDI deployment) - the project is many times larger than yours.

All parameters of the virtualization system (ups, raid, smart, temperature, hypervisor load, VM health) need to be monitored with email notification

Everything is realizable.

Downtime allowed for guest VMs

I propose to design systems immediately without such deplorable moods.

The main thing is the safety and recoverability of guest VMs

using the Vaaem solution is the best solution for reserving a Vmware / Hyper-V virtual infrastructure The
question, as I understand it, is only in storage.
If you want to use part of the storage on each server, then you will not make a full-fledged cluster (with the exception of using the VMware vSphere Storage Appliance), but there is a minus. You all have 4 disks on the host, if these disks are still driven into raid, then the total IOPS performance will not exceed 140 IOPS (taking into account 30% cache and r / w 70/30).
If you make (buy) storage for 12 disks, then the performance of the disk system will increase by 4.5 times (calculation of raid 10 12 disks against raid 5 4 disks). This will resolve the issue of the total moon. For migration, allocate a separate network card, a card for accessing storage (icsci).
If there is a question of comparing hypervisors, the next edition of the Gartner Magic Quadrant on x86 Server Virtualization 2013 speaks for itself

. And yet, it’s not difficult to delve into and learn vmware.
I will be glad if calculations or comments help you

Answer 4

Good afternoon!
Look at virtualization from RedHat, it seems to me there is a lot of interesting things there. There is a commercial version (RHEV) for virtualization of servers and desktops (from your question it is not entirely clear what you want to virtualize), today it already works quite well.
You can also see a free analogue (on the basis of which a commercial distribution kit is then made, as I understand it): www.ovirt.org/Home Perhaps a free analogue will be enough for your tasks, you can implement it in a couple of days.
I hope the information will be useful to you.

Answer 5

I will support the previous speaker about Proxmox. However, not everything is so rosy with him. But in general, it does not cause much trouble (except for hanging backups somewhere once a month).

Answer 6

What to choose for independent implementation? Please share your real experience.

What's wrong with vSphere Essentials Plus? If you already have a purchased license for it, why look towards other solutions?

Answer 7

The Vmware vSphere Essentials Plus Kit license is just for three servers, what you need. If you are not going to renew your subscription to VmWAre, then as Proxmox advised above.

Answer 8

All parameters of the virtualization system (ups, raid, smart, temperature, hypervisor load, VM health) need to be monitored with email notification

Do it yourself.
For SMART monitoring - smartmontools. RAID - appropriate software for your controller + scripts. Control of availability and parameters of the host and guest systems - Zabbix and analogues.
If your admin knows Debian well, then all this should be no problem for him.

The virtualization system must be able to gracefully shut down before the UPS is completely discharged

Again - do it yourself. It's not the job of the virtualization system to keep track of your hardware.
I would recommend the Xen Cloud Platform (based on CentOS), which is in no way inferior to the commercial Citrix XenServer ( comparison ). You can manage everything through the convenient Citrix XenCenter interface or analogues. You can install from scratch / use an existing Debian and add XAPI toolstack to it , get all the XCP features in your familiar environment.
VMware ESXi, after many years of experience with Xen, absolutely did not like it.