Destination unreachable on some devices on a VLAN

I have a core switch, model GSM7328FS that I am having a bear of a time with.

I have 2 ports on assigned to VLAN 3... PVID set to 3 on those 2 ports. I have a network closet and remote switch at the end of each cable run for those 2 ports. The 2 remote switches are assigned IP 172.22.6.2 and 172.22.6.3 respectively. The core switch is assigned 172.22.6.1 as the VLAN 3 interface and 172.22.4.15 as its primary interface on the primary VLAN.

Everything on the first port where switch 172.22.6.2 is located is working absolutely fine. Certain things on the second port where switch 172.22.6.3 is located is flaky as hell.

Pinging 172.22.6.3 from the 172.22.4.0 network comes and goes. Some print servers on the 172.22.6.0 network are coming and going. A reboot of the devices that have "gone" will typically fix them for a short period. PC's behind the 172.22.6.3 switch seem to be much more stable and don't seem to drop off the network, BUT large file copies from those devices won't complete even though they never stop pinging.

When 172.22.6.3 has stopped being reachable from the 172.22.4.0 subnet I can remote control one of the PC's on the 172.22.6.3 subnet and log into the 172.22.6.3 switch even though I can't do it from any IP on the 172.22.4.0 subnet. Sometimes when I log in from a machine on that subnet, it triggers the switch to become responsive to the 172.22.4.0 subnet again for a period of time.

When one of the devices behind 172.22.6.3 becomes unresponsive, I can do a tracert from a windows PC on 172.22.4.0 network and I get "172.22.4.15 reports: Destination host unreachable".

If I look at the address table on 172.22.4.15, the MAC address for 172.22.6.3 is there and showing on the correct port even though I can't ping it. When it is in this state, pinging 172.22.6.3 directly from the 172.22.4.15 switch also fails.

Please help me!