On Mon, 26 Jan 2015, Michael Dukelsky wrote to mark lewis:

 MD> [...skipped...]
 ml> do you think that something like this might be able to be done for
 ml> other OSes?? and also for when operating as a seperate task? some time
 ml> back there was asked for semaphores which could be used to tell binkd
 ml> to freeze and close all its files when the current session(s) were
 ml> completed and it would then create a frozen semaphore signalling that
 ml> it was frozen and external maint tasks could be performed... then the
 ml> removal of the original freeze semaphore would signal binkd to return
 ml> to operation and remove its frozen semaphore that it had created...
 ml> these semaphores would be disk-based zero byte files in most cases...

 MD> As to me I don't think that all this is necessary. I have an init
 MD> script for binkd in CentOS 5, so
 MD> /etc/init.d/binkd stop and /etc/init.d/binkd start
 MD> do the job for me. But Pavel who is the binkd maintainer may have a
 MD> different opinion.

initing binkd is not where the desired function lays... being able to pause a
running binkd for (eg:) log rotation is the desire... not all systems running
binkd operate like winwhatever or *nix... the main goal, though, is to freeze
binkd so data is not lost when (eg:) rotating the logs as a new session might
start during the copy process... it depends on how a system does its
rotation... on some systems the operator has to craft such rotation tasks by
their hand in their own scripts...

eg: for log rotation on one of my systems, i /have/ to shut down binkd, do the
rotation, restart binkd... this is not optimal because then binkd unloads
everything and reloads everything as well as altering its position in the
running tasks list... this is undesirable and thus the original request several
years ago and the continuation of this one now...

and in case i have forgotten in previously, thank you very much for your
responses and the work you guys do in binkd :) 

)\/(ark