|Date: ||Thu, 18 Apr 2002 08:41:19 -0700|
|Reply-To: ||MarK Dodd <markdodd9999@HOTMAIL.COM>|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|From: ||MarK Dodd <markdodd9999@HOTMAIL.COM>|
|Subject: ||Re: locking problem|
|Content-Type: ||text/plain; charset=ISO-8859-1|
Apologies ... I should have put a bit of detail about what I was
actually doing when the lock occured.
The program is completely a batch program and is the only program
runnign at the time (dependencies have been set to ensure this), and
no interactive users are on the system that time of day - so no
var/fsp/etc windows (and BTW no users access this file anyway). We
use BJS as our batch controller.
The rough datastep is :
%do z=1 %to &_n;
%if %sysfunc(exist(mydata.u&&date&z)) %then %do;
other data step statements ... ;
note : Ive left off the code that sets up &&date&z ... they are of the
value of putted sas dates in date7 format (27mar02).
It works for all the rest of the days (16mar02-17apr02). Most days
this program runs with out problems, but *sometimes* it fails. this
dataset existed before the program ran. Its nothing terribly
complicated (there are no recursive call executes, which conceptually
might be able to cause this sort of thing <shrug>)
thanks to the people who emailed me suggestions - just to let you know
stuff I am following up :
1) backup software. It should have been running at 10am, so this is
unlikely to be the cause, but I am checking.
again cheers for the help.
ghellrieg@T-ONLINE.DE (Gerhard Hellriegel) wrote in message news:<200204181024.g3IAO9818011@listserv.cc.uga.edu>...
> On Thu, 18 Apr 2002 02:30:04 -0700, MarK Dodd <markdodd9999@HOTMAIL.COM>
> >Im using sas 8.2 on NT4 (dont know what serv pack). I get the
> >following error-
> >"ERROR: A lock is not available for MYDATA.U27MAR02.DATA, lock held by
> >another process"
> >There are no other sas processes running at the time this process runs
> >(2am), and there are no users on the system. the error seems to occur
> >randomly through this job about once every few weeks. The process
> >loops through a number of days in the month and at this point in the
> >job, has finished about a dozen days in the loop.
> >The sas tech site suggests virus software may be a problem, but it
> >runs once/week and that wasnt last night.
> >Any ideas ???
> Hard to say anything, without knowing something about your program.
> Something like that you get for example, if you have opened a var / fsp /
> whatever-window and then submit a program what tries to access the related
> dataset. That might be something, what your SAS program does. But to have a
> concrete idea it is necessary to know that.