single 0.0.2

Ensure single instance of a command without leaving stale lock files.

A wrapper command to provide an advisory lock, without leaving stale lock files. Usage: single.py [options] -c command args..

Options:

-h, --help

show this help message and exit

-f LOCK_FILE, --lock-file=LOCK_FILE

Path to the lock file. Default is provided based on the command path if omitted.

-s, --status

Check to see if the file is locked, and if so, by which process. Exit status is 0 if unlocked, 1 if locked.

flock(2)’ed file is used to provide locking. This will not leave stale lock files around. Note that the existance of lock file does not mean that there is an outstanding lock; exclusive flock must be held by a process. So use ‘single.py –status -f /tmp/foo.lock’ to see if the lock is held.

The OS releases the lock upon process termination, so the lock file is released regardless of how the job terminated.

Invocations:

• single.py -c long-running-scrit arg1 arg2 will ensure only one long-running-scrit will run at a time. Default lock file, specific to the command, is used in the absence of -f option.

• Lockfile can be explicitely specified as: single.py -f /tmp/lrs-foo.lock -c long-running-scrit foo Two jobs using the same command could be run concurrently by using different lock files, like: single.py -f /tmp/lrs-bar.lock -c long-running-scrit bar

• Use –status (-s) option to check if a command or a file is locked: single.py -s -f /tmp/foo.lock single.py -s -c long-running-scrit

Example: $ single.py -c sleep 3 & for x in {0..6}; do single.py -s -c sleep; sleep 1; done [1] 32567 locked by 32567: /tmp/single.py_bin_sleep.flock locked by 32567: /tmp/single.py_bin_sleep.flock locked by 32567: /tmp/single.py_bin_sleep.flock [1]+ Done single.py sleep 5 not locked: /tmp/single.py_bin_sleep.flock not locked: /tmp/single.py_bin_sleep.flock