Howdy folks,
I've recently been having problems with my Ubuntu VPS on my Windows server. I have access to both the Host node and the guest node (since this is a multi-use server and I wanted to use both Windows and Linux services), and I guess I thought I set it up properly but it seems there's a networking problem on the server.
There's periods when the Ubuntu VM has major packet loss while the actual host server is mostly fine. Sometimes when I'm rsyncing large files (12 GB worth of files), the network gets dropped on the entire Ubuntu VM and I get the error "ping: sendmsg: No buffer space available". This was reproducible whenever I'd rsync out larger files.
Basic research shows this is usually due to an incompatibility of the on-board NIC on the Server (of course this is assuming Ubuntu was installed on the base hardware, not within a VM). I'm assuming this is a problem with my bridge setup.
Anyways, this is a Hyper-V Generation 1 VM. The bridge is basically all default variables requested when setting it up through the Wizard available on the Windows Server 2012 R2 (Datacenter Edition). Upon further Google Searching, I've found that we have to increase the memory allocation for the packets, and was suggested I add the following lines into sysctl:
net.core.rmem_max=16777216
net.core.wmem_max=16777216
net.ipv4.tcp_rmem=4096 87380 16777216
net.ipv4.tcp_wmem=4096 65536 16777216
In addition, the following command was executed: echo 83886080 > /proc/sys/net/core/wmem_max
Anyways, even with these changes I'm still having some problems. Anyone know of a good way to start troubleshooting this issue?
Thanks!
I've recently been having problems with my Ubuntu VPS on my Windows server. I have access to both the Host node and the guest node (since this is a multi-use server and I wanted to use both Windows and Linux services), and I guess I thought I set it up properly but it seems there's a networking problem on the server.
There's periods when the Ubuntu VM has major packet loss while the actual host server is mostly fine. Sometimes when I'm rsyncing large files (12 GB worth of files), the network gets dropped on the entire Ubuntu VM and I get the error "ping: sendmsg: No buffer space available". This was reproducible whenever I'd rsync out larger files.
Basic research shows this is usually due to an incompatibility of the on-board NIC on the Server (of course this is assuming Ubuntu was installed on the base hardware, not within a VM). I'm assuming this is a problem with my bridge setup.
Anyways, this is a Hyper-V Generation 1 VM. The bridge is basically all default variables requested when setting it up through the Wizard available on the Windows Server 2012 R2 (Datacenter Edition). Upon further Google Searching, I've found that we have to increase the memory allocation for the packets, and was suggested I add the following lines into sysctl:
net.core.rmem_max=16777216
net.core.wmem_max=16777216
net.ipv4.tcp_rmem=4096 87380 16777216
net.ipv4.tcp_wmem=4096 65536 16777216
In addition, the following command was executed: echo 83886080 > /proc/sys/net/core/wmem_max
Anyways, even with these changes I'm still having some problems. Anyone know of a good way to start troubleshooting this issue?
Thanks!