cPanel + Hyper-V = slow network ?

Senor

• January 23, 2017 14:14

Hello all - I recently migrated a server from an older CentOS box to a Hyper-V install on CentOS 7 (minimal install per the install guide). Since the move, I've noticed ridiculously slow network speeds even between the VM and the host. Everything works but it's very slow. These are Dell T320 servers running Windows Server 2012 Standard R2, dual 10Gb Broadcom NICs and more than enough CPU power and RAM to run a very basic cPanel install. Pings to the VM fluctuate between 0 and 200ms, though there is no packetloss. Telnet'ing into the VM on standard ports (like 25 for SMTP) takes few seconds before the server responds. Websites load much slower than before and I can see pages actually render (long drop down lists, for example), whereas they were instant prior to the move. Googling shows it might be a Virtual Machine Queue issue, but the problem remains with it enabled and disabled. I've updated the drivers to my NICs and even tried some registry entry changes all to no avail. I've also swapped virtual switches assigned to the cPanel VM with no luck. Other VMs (including another Linux VM) have no network issues. I want to believe this is an issue with my Hyper-V host, but I cannot duplicate the network problems with any other server. Any ideas on things I could check (both Hyper-V host and the VM itself)? Thanks in advance.

• 

  cPanelMichael

  • January 26, 2017 17:44

  Hello, It looks like ticket number 8153007 was opened to investigate this issue. Feel free to let us know the outcome once the ticket is closed. I'll also leave this thread open for additional user-feedback from anyone else that may have experienced a similar issue. Thanks!

  0
• 

  Senor

  • January 26, 2017 17:48

  Hi Michael - Thanks for the update. It turns out it was a VMQ issue. Why the original toggling of the feature didn't work, I don't know, but enabling and disabling it again seems to have fixed the issue. You can probably lock this thread. Thanks!

  0
• 

  cPanelMichael

  • January 26, 2017 17:51

  I'm happy to see the issue is now resolved. Thank you for updating us with the outcome.

  0
• 

  digitaliway

  • November 05, 2023 18:37

  I know this thread is really old, but I am having a similar issue that no one seems to be able to figure out. we migrated from a dedicated server running cloud Linux 7, litespeed, CPanel 110.0.14 TO a VM CloudLinux v8.8.0, litespeed, CPanel 110.0.14. The new cloud linux VM is on top of windows server 2022 hyperV and has 20 CPU cores allocated and 90GB ram all on SSD drives. PHP processes seem to delay or have a strange pause for a half second then process causing website delays (this did not happen on dedicated). we have a constant high server load in the averages of 6 (was 1.5 on dedicated) and on every reboot using WHM the VM will become almost unresponsive with LSphp processes piling up in the process manager and when LSphp processes piles up then MYSQL processes pile up which ends up in an almost crash state of the VM. I am leaning toward a VM config issue but have no idea where to look or how to troubleshoot this. If it is a VM config issues I need to know what to tell the datacenter specifically do they can engage with it. I have tickets open with cpanel, cloud linux, litespeed and some independent techs looking at this and no one can figure this out so far. on the VM I do have a shmem-rss display in the screen shot showing "memory cgroup out of memory" . any help or direction would be appreciated.

  0

Please sign in to leave a comment.