HP Tru64 UNIX and TruCluster Server Version 5.1B-6 Patch Summary and Release Notes

• Fixes a problem in which questionable behavior by non-TRU64 NFS clients results in NFS
over TCP server threads hanging.
• Fixes a problem related to the deletion of binding cache entries in mobile IPv6, which affects
multi-RAD systems like the AlphaServer GS1280.
• Addresses issues that may be seen as panics or hangs having to do with the UBC.
• Introduces the vm_overflow tunable.
• Provides enhancements for NetRAIN operations.
• Corrects race in streams I/O completion and timeout logic.
• Adds the ability to decouple RPC client retransmission activity from UDP or interface
transmission completion handling.
• Fixes a condition that causes a connection timeout error when using socket keepalive options
with small values on sockets over a localhost connection.
• Corrects the cause of the a panic: "panic: Hashing that is not marked VPP_REPL."
• Corrects a problem in the kernel dli module that causes system hangs and crashes on
multi-CPU systems. The crashes occur with the panic string "lock_terminate: lock held".
• Fixes a logic error in the virtual memory macros used for accessing the memory descriptor,
based on physical address.
• Corrects the panic “trap: invalid memory read access from kernel mode” in pshared code.
• Corrects a problem with the randomization of initial TCP sequence numbers.
• Allows multiple retries to mount the root file system and the ability to adjust the retry period.
• Enhances the range of generic sysconfig parameters and eliminates unnecessary sysconfig
warning messages.
• Corrects several problems related to use of the ifconfig [interface] IPv6 up command, where
interface is either cluster interconnect, ics0, or configured v6-in-v6 tunnel.
• Fixes an NFS client kernel loop that usually results in a lock time limit exceeded panic on
partitioned GS1280 class systems in partitions where CPUs have different numbers of
neighbors. That usually requires a partition of more than four CPUs.
• Ensures that the socket send buffer size will be twice the maximum segment size if the mss
is 1460 bytes.
• Fixes a problem that causes a hang or crash when auditing the network events
“net_tcp_stray_packet,” “ net_udp_stray_packet,” and “net_tcp_rejected_conn.”
• Corrects a problem with the audit object deselection mode in which it was possible for events
that modify a target object to be deselected, resulting in no audit record being generated.
• Corrects for a leak of CONTROL mbufs.
• Fixes a problem in which a threaded program tries to wake up a thread that is already
running.
• Enhances the fuser command to provide a cluster-wide query capability. A revised fuser(8)
reference page describes this enhancement.
• Fixes a problem in which the traceroute command will timeout when probing the cluster
alias address from a cluster member.
• Corrects a memory leak against the 128-byte bucket when rad_gh_regions are configured
and nshmget() is used.
• Adds support for ifconfig [inet6] delete [abort], for IPv6 connections.
• Corrects a problem in which the netstat -s command does not display the correct global
statistics on multi-CPU systems if lower-numbered CPU slots are empty.
• Corrects a memory leak in IPv6 transmission code.
• Corrects a condition that results in a failure to release file region locks (NFS client only).
• Corrects a panic “vm_pg_free: page on o/h list.”
• Fixes a process deadlock that occurs when rename is called with “.” as the target.
Summary of Base Operating System Patches 101