lockd(8) CLIX lockd(8)
NAME
lockd - Runs the NFS network lock daemon
SYNOPSIS
/etc/lockd [-t timeout] [-g graceperiod] [-h hashsize] [-l k2utimeout]
FLAGS
-t timeout Specifies the number of seconds as the interval instead
of the default value (15 seconds) to retransmit lock
request to the remote server.
-g graceperiod Specifies the number of seconds as the grace period
duration instead of the default value (45 seconds).
-h hashsize Specifies the number of hash buckets instead of the
default (29).
-l k2utimeout Specifies the number of seconds as the interval instead
of the default value of 2 seconds to retransmit kernel
lock manager requests. This is the timeout value used
for local lock requests.
DESCRIPTION
The lockd command processes lock requests sent locally by the kernel or
remotely by another lock daemon. The lockd command forwards lock requests
for remote data to the server site's lock daemon through the RPC/XDR
package. Then lockd requests the status monitor daemon, statd, for
monitor service. The reply to the lock request is not sent to the kernel
until the status daemon and the server site's lock daemon have replied.
If either the status monitor or server site's lock daemon is unavailable,
the reply to a lock request for remote data is delayed until all daemons
become available.
When a server recovers, it waits for a grace period for all client site
lockd processes to submit reclaim requests. Client site lockd processes,
on the other hand, are notified by the statd of the server recovery and
promptly resubmit previously granted lock requests. If a lockd fails to
secure a previously granted lock at the server site, the lockd sends
SIGUSR2 to a process.
EXAMPLES
To start lockd, specifying a 20-second grace period and a 10-second
interval for retransmitting lock requests to remote servers, enter:
lockd -g 20 -t 10
2/94 - Intergraph Corporation 1
lockd(8) CLIX lockd(8)
FILES
/dev/dsk/* Disk device files.
NOTES
The lockd command should be invoked early during the transition from
single user to multiuser, so that no other processes have the opportunity
to get a standard System V lock. If there are active or sleeping locks in
the standard System V record locking code, lockd attempts to migrate them
to the user level process. This is done so that the active locking
processes will not have their locks destroyed by the starting lock
manager.
With the ISS 06.00 release of NFS, PC and UNIX locks are merged. A PC is
allowed to lock a file only after it has secured both the cooresponding
UNIX lock (granted by the kernel on the server) and the PC lock. For
example, if a UNIX machine has locked a file, a PC will not be allowed to
lock the same file. The mapping between PC and UNIX locks is as follows:
PC lock UNIX lock
record lock WRITE lock
record unlock UNLOCK
DENY_WR file sharing mode WRITE lock entire file
DENY_RDWR file sharing mode WRITE lock entire file
DENY_NONE file sharing mode UNLOCK entire file
DENY_RD file sharing mode (no cooresponding UNIX lock)
DIAGNOSTICS
Note that all diagnostics are displayed on the system console.
lockd: main(): /dev/lockd open failed
The /dev/lockd file could not be opened because the user lacks
superuser privileges or another lockd command is already executing.
EXIT VALUES
The command has the following exit codes:
0 Upon receipt of a signal, lockd was able to shutdown normally,
passing control of local locks back to the kernel.
-1 Upon receipt of a signal, lockd was unable to shutdown normally.
RELATED INFORMATION
Commands: statd(8)
Functions: fcntl(2), signal(2)
2 Intergraph Corporation - 2/94