Problem
The following error messages may be seen on the console or in the syslog file:
WARNING: NLM: RPC call failed: RPC error: RPC_CANTRECV, errno 22 WARNING: NLM: RPC call failed: RPC error: RPC_PROGUNAVAIL, errno 118 WARNING: NLM: no memory for client's address in nlm_blklock, lock request dropped WARNING: NLM: unlock request to server <IP> failed: RPC_CANTRECV
Shortly thereafter, the system may panic with the message:
PANIC: k_trap - Kernel mode trap type 0x0000000ECrash dumps indicate panics in klm_lock and in nlm_runblklcks.
Cause
There are actually two related problems happening here. The first
is a condition whereby lockd fails on a server and attempts to
clean up locks that reference NULL file table entries. The second
is a situation in which memory for a lock is allocated to a NULL
pointer when the referenced lock address becomes NULL. Solution
Install Support Level Supplement (SLS) OSS449F, the Network Performance SLS for SCO OpenServer 5.0.0, 5.0.2, and SCO Internet FastStart Release 1.0.0, 1.1.0.
See also
TA107704, "What is the SLS OSS449F Network Maintenance Supplement?"