Hi, I recently set up a virtual dev environment on ESXi composes of a vSRX cluster (2 vCPU, 8G memory each instance) and a pair of vMXes, vMXes act as Internet edge while vSRX cluster performs firewall function, I also have a physical dev environment in parallel on the same rack sharing the same Internet connection, physical SRXes are SRX-100es, a simple speed test on hosts behind vSRX/SRX-100 reveals that the throughput across vSRX/vMX is only ~10% of the throughput over physical SRX (200Kbps vs 2Mbps, why only 2Mbps on SRX-100 is irrelavent ), this is way lower than a vSRX/vMX should be able to handle, where should I look to find the root cause?
There could be some many things to check, to start with:
1: What is the version of the vSRX and the type of virtual interface, e.g. VMXNET 3 etc.
2: What is the interface status speed etc, any errors, drop , pkt in/out in the 'show interface extensive'
3: show pfe statistics traffic https://www.juniper.net/documentation/en_US/junos/topics/reference/command-summary/show-pfe-statistics-traffic.html
4: Session details to see the WSF etc are negotiated properly.
5: Try to use only vSRX no other device , and perform your test to see if there is any improvement
6: CPU/memory status during the test 'show chassis routing/forwarding'
Are all the virtual devices properly licensed?
I believe they are bandwidth limited until the license is applied.
Thanks, it turns out that vMX in front of vSRX did not have the trial license installed.