Venet use in container takes several seconds to establish outbound network connection

[mastermindpro]

I've been using PVE 1.5 on a few hosts running a mix of KVM and OpenVZ guests. Rather suddenly, one of my OpenVZ guests that uses only venet interfaces is having network connectivity issues. Outbound connections all complete, but each session takes several seconds to start transferring data. There is no delay at all for inbound connections. I've tried restarting the container, removing assigned IP's and re-adding them, but nothing seems to help. This is the output of pveversion -v on the host:

pve-manager: 1.5-8 (pve-manager/1.5/4674)
running kernel: 2.6.24-10-pve
proxmox-ve-2.6.24: 1.5-21
pve-kernel-2.6.24-7-pve: 2.6.24-11
pve-kernel-2.6.24-10-pve: 2.6.24-21
pve-kernel-2.6.24-9-pve: 2.6.24-18
qemu-server: 1.1-11
pve-firmware: 1.0-3
libpve-storage-perl: 1.0-10
vncterm: 0.9-2
vzctl: 3.0.23-1pve8
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.11.1-2

Anything I can try to diagnose what the problem(s) is/are?

 

[mastermindpro]

Re: Venet use in container takes several seconds to establish outbound network connec

I fixed my own problem, sort of. The container in question was migrated from a non-Proxmox OpenVZ node, and it still had the old tcpsndbuf limits. Now that the container is being used more, these limits were hit sooner. I increased those and restarted the container to find that, for some reason, the container couldn't connect to the public IP of our DNS servers. I changed it to the private IP, and it worked fine.

The first problem makes sense, but the second one doesn't. Oh well. Sorry for the noise.