GitHub user fnavidan edited a discussion: High Latency Between Two VMs In a L2 Private Guest Network
Hi everyone, I am running Apache CloudStack v4.22.1.0. I have created a private L2 guest network, then added a second NIC to each VM connected to private network. When I ping other VMs in the same private network, I get following results: [root@da1 ~]# ping -c 100 10.0.0.2 PING 10.0.0.2 (10.0.0.2) 56(84) bytes of data. ... 100 packets transmitted, 100 received, 0% packet loss, time 100024ms rtt min/avg/max/mdev = 0.266/2.122/11.104/2.249 ms [root@da1 ~]# traceroute 10.0.0.2 traceroute to 10.0.0.2 (10.0.0.2), 30 hops max, 60 byte packets 1 da-node2 (10.0.0.2) 1.334 ms !X 1.088 ms !X 1.013 ms !X All VMs are running on the same KVM hosts which should normally give under 1ms ping rtt between guests but it doesn't. My test scenario implements GlusterFS having 2 replicas + 1 arbiter which is sensitive to latency. Currently when I test write performance on the gluster volume, I get very low throughput around 30MB/s while writing to the local filesystem of VMs gives me a 350-400MB/s write speed. Additionally, all VMs are running on the same primary storage which is mounted as a local NFS on the KVM host, and all nodes connected to the same l2 private network. As my gluster traffic is passing through L2 network, there must not be any extra networking overheads in a l2 network type. GitHub link: https://github.com/apache/cloudstack/discussions/14380 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected]
