NethServer Version: 7.2009
Module: Core
Is anyone else running into an issue when running NethServer on a VM where pretty much everything is ungodly slow and pretty much unusable?
Active Directory services seem to work smoothly, but things like NextCloud, VPN, SOGo, Email, and the File Server functions are basically unusable. I’m wondering if this is just me or if this is a known issue.
Hi
I’m running about 30 NethServers for clients and myself, all now running on Proxmox, and faster than the hardware they replaced. Some migrated from VMWare ESXi (I was a very longtime VMWare user, from before 2000!).
My 2 cents
Andy
I’m running an instance on VMware ESXi with 16GB of Ram, and 1TB allocated and everything except AD services are extremely slow. Any recommendations, things to check?
Hi
Do you have VMWares’s promoscious mode active on the NIC your NethServer uses?
AD on Nethserver uses a Container/Jail, with it’s own IP. That’s why promoscious mode is needed.
Make in Nethserver (Must be your internal DNS) a DNS entry pointing to the IP of your AD, eg:
ad.mydomain.com 192.168.20.11 (or whatever you’re using…).
AD Clients should (must) use NethServers DNS.
16 GB RAM is more than enough, I have clients at three sites with one NethServer, a total of 35 users, Mail, Nextcloud, AD, Zabbix - with 12 GB RAM allocated in Proxmox!
If you have a spare machine to play around, set up Proxmox and check it out.
I think you’ll love it. All usable for free, including full HA cluster capabilities, live Backups of Windows and other OS included… A lot of the people here use Proxmox underneath.
Hope this helps.
My 2 cents
Andy
Thanks for everyone’s insights. So just for giggles, I ended up moving our instance of NS to a physical machine just to rule out any issues with the VM causing latency or a delay and I’m seeing the same issues!
VPN traffic is ungodly slow, loading anything from NS whether it be the GUI, VPN, NextCloud, SOGo, is super slow, but being on the network itself everything there doesn’t appear to be anything wrong with the flow of network traffic, devices can quickly and easily authenticate through the directory, and pings are less than 1ms, so it doesn’t look like it’s an issue with the physical hardware.
I know it’s not the connection or the switch itself, because other servers on the same switch load almost instantly with no latency or degradation issues.
Any insights on what could be causing issues?
Check if the physical host does not have too much load of disk IO, this could be a bottleneck.
Thanks @stephdl, there doesn’t really appear to be any load when checking the disk IO and attempting to load any of these features. It’s minimally effected.
What resources did you allocated to the VM ?
16GB of Ram, 1TB Disk, and a 4 sockets.
But it’s on a physical machine now and I’m seeing the same thing.
A physical host like this and nethserver is not skyrocketing, I suspect a hardware problem or a hardware not compatible with centos7.
I’m not sure if that would be the case, this server is brand new, so hardware itself I would think would be fine on it, additionally we have three other hosts with the same exact configuration and they’re ungodly fast. Again it just seems to be anything external or outside the network NethServer is on. I’m wondering if maybe there is something hindering the connection I’m not aware of.
If it runs at the same hardware as the virtualization before, I would try to test it on another hardware.
I would imagine it wouldn’t be an incompatibility with the hardware because I have deployed about 4 of these on this exact same configuration, and it’s only this one instance which is having the slow down.
Additionally it’s only effecting the GUI, NextCloud, SOGo externally. If it was a hardware issue, I’m assuming it would be effecting it both on the LAN and WAN. And not just the WAN side of things.
Hi
In IT it’s not unusual to have DOAs (Dead on Arrival). January 2020, we migrated a Doctors from SME Server to NethServer, and PCs from Win7 to Win10. 5 new Intel NUCs…
We installed Win10, created a master image, and loaded this image on the 4 other NUCs. All 4 worked at the start, only one had “funny” issues. If rebooted, it lost contact to the AD, sometimes Win10 thought it was freshly installed. The supplier changed Board, Casing and CPU / RAM twice, before giving us a new NUC replacement. This finally worked!
You can - for example - fire up that server with a LIVE Distro, just to verify it might have hardware issues, which the other 3 don’t have!
My 2 cents
Andy
check logs
messages
firewall.log
secure (?)
httpd/{*error_log,*access_log}
Try to disable shorewall at all for a test : shorewall clear (shorewalll start)
I have no idea why, but clearing the firewall appears to fix the issue, so there is something on the firewall that is creating the issue. When I start it again it slows down to a crawl.
So something somewhere is blocking or slowing down the connections.
check firewall.log, everything is logged, check also what configuration you did on the firewall
does the NIC is exactly the same ???
The configuration for the firewall was left as default. Nothing has been changed. We let Nethserver adjust the firewall. I don’t know what you mean by “does the NIC is exactly the same???”
is it a clone you did or you install the OS from the CD, sometimes it could bring issues, maybe the network interface is not the exactly the same
The interface is the same as the previous server (VM). Like I said, when I enter the command “shorewall clear”, everything is lightening fast. However when I use “shorewall start”. It slows down again.
Is there a way to force Nethserver to use another firewall system, say like CSF?