I've seen this issue on a few of our Dell PE T620's with QLE 2462's installed. The issue stems from an interrupt bug in the Linux kernel. The resolution is/was disable interrupts for qlogic card:
VMware KB: I/O activity pauses on virtual machines with QLogic 81xx series CNA cards on ESXi 5.x
And then install the VMWare latest driver/firmware for the Qlogic card (can obtain from your OEM provider )
I found the interrupt disable only worked with the new VMWare driver.
IF you look in your VMKernel logs, you should see the Firmware restarting and a panic state with APD enabled. At this point, there is no restarting. VMWare is protecting the HBA from receiving bad, out-of-sync, or corrupted I/O commands. You can increase the APD timeout, but you're not really correcting the problem... VMWare specifically put the APD timeout in as a safeguard so that you halt all I/O before re-attaching.