Hello!
That is very strange indeed, but still all the symptoms points us to a hypothesys that traffic to outside subnets is not routed by the servers to VLAN-interfaces, as unlike many other vendors, IP routing is enabled in all Comware products by default, so there should be nothing that stops it.
Could you check the following:
1. Enable IP TTL expired and IP Unreachable ICMP messages in the switch:
system-view
ip ttl-expires enabled
ip unreachables enabled
return
2. run traceroute from a host in one VLAN to a host in other VLAN
3. run 'ipconfig /all', 'arp -a' and 'route print' on both hosts (valid for Windows hosts, if you use other OS, collect IP information of all network interfaces, routing table and ARP cache)
4. From the 5130 ping both servers using different source VLAN interface, for example if you have two VLANs - 10 and 20, with ip addresses assigned 10.0.10.1/24 and 10.0.20.1/24 respectively, and your servers in each VLAN are 10.0.10.100/24 and 10.0.20.100/24 (for example) run two pairs of ping commands:
a. ping servers using source IP of their VLAN interface:
ping 10.0.10.100
ping 10.0.20.100
b. ping servers using other VLAN as a source:
ping -a 10.0.20.1 10.0.10.100
ping -a 10.0.10.1 10.0.20.100
If you will run a Wireshark on each host at that time, it will greatly help us as we will see if ICMP echos are reaching the servers and if servers reply at all, and if they do - what is the destination MAC address of these replies - it should be equal to 5130's VLAN interface MAC (can be checked by 'display interface <Vlan-interface> | i hardware')
5. Generate 'display diag' after all tests above.
Then share the information collected on Steps 2-5 with me - console outputs, Wiresharek traces, diag from the 5130, hopefully together we'll find the root cause.