854 Subj : shared COMM port To : Elwyn Goldsbury From : Mike Tripp Date : Mon Jan 22 2001 05:38 am Hello Elwyn. 22 Jan 01 19:12, Elwyn Goldsbury wrote to Mark Lewis: EG> thanks Mark - thats the way I may need to go - currently I am using EG> two modems - but I would rather have a Cron triggured event that EG> shuts down binkley does the internet connection them fires up binkley EG> again - just looking for anyone who may hav put some code together to EG> do it !! I had a similar requirement to be able to knock Binkley offline long enough for a maintenance routine to trim its logs. Because my processes were scattered between the DOS mailer node and and OS/2 node across a Netware LAN, the solution required a combination of Binkley's flag file support (which can be used to force Bink to exit with a specified errorlevel) and a semaphore file written to the server. The logic flows something like: Log process creates flag file to force Bink to exit and then loops, waiting for the appearance of a semaphore from the Bink process to acknowledge that it indeed has exited. Binkley process finishes doing whatever work it might have been doing when the flag was created, detects the flag, exits with specified errorlevel, and branches to BAT/CMD logic that creates the semaphore and loops while that semaphore exists. Log process exits wait-for-semaphore loop, does its work, and then deletes semaphore. Binkley exits wait-while-semaphore loop and recycles. Sounds convoluted, but it really only involves an additional 6 or 8 lines scattered across the two batch jobs to implement. You'll spend more time in the Bink docs figuring out exactly how that damn flag file has to be named to match your Bink implementation. For mine, it is something like BTEXITxx.yy: where xx is the task number (node) in hex and yy is the errorlevel in hex...but this may vary between versions or in XE releases, so check your dox (including WHATSNEW) before you start. ..\\ike --- GoldED/2 2.50+ * Origin: TechKnowledgy at Work (1:382/61.1) . 0