...
peer_timeout can be increased to fit in more re-transmits or LND transmit timeout can be decreased.
Shadow OW: Let's not use a nickname Alexey Lyashkov made a presentation at LAD 16 that outlines the best values for all Lustre timeouts. It can be accessed here.
...
MD is always protected by the lnet_res_lock, which is CPT specific.
Other data structures such as the_lnet.ln_msg_containers, peer_ni, local ni, etc are protected by the the lnet_net_lock.
The MD should be kept intact during the resend procedure. If there is a failure to resend then the MD should be released and message memory freed.
...