Re: Package automation in /etc/daily

[Steven Bellovin <smb%cs.columbia.edu@localhost>]

On Jan 17, 2010, at 11:53 AM, Julio Merino wrote:

> On Sun, Jan 17, 2010 at 4:46 PM, David Laight <david%l8s.co.uk@localhost> 
> wrote:
>> On Sun, Jan 17, 2010 at 05:42:35PM +0100, Bernd Ernesti wrote:
>>> * it can cause more workload on ftp.netbsd.org and exceed the connection
>>>   limit on ftp.netbsd.org if a lot of systems use it
>> 
>> Actually, since all the systems in a timezone will connect at
>> roughly the same time (2am local time), the whole thing is
>> probably a very bad idea!
>> 
>> In fact, since the list probably changes infrequently, a mailing list
>> update scheme would be more appropriate.
> 
> Except that we don't have this now and we have been using and
> advocating a different scheme, so I'm not planning to change it.  I'm
> just trying to make things easier for initial configuration of a
> machine with packages installed.
> 
> We have had audit-packages for ages, and we have encouraged everyone
> to use it.  Even more, we have even provided sample crontab entries,
> showing the exact same scheduled time for everyone.
> 
Maybe the default cron entry should be prefixed with (for ksh)

        sleep $((RANDOM%30)) 

or some such.  (I'm 100% serious about a random delay, though I'd most likely 
suggest building it in to the fetch command or script, activated by an option 
used by the cron invocation.  If we really wanted to be kinky, we could make 
the option only active if the program were running without a control tty, i.e., 
from cron instead of by the administrator.  That's worth serious consideration, 
given how often some people (including me) just cut-and-paste such things from 
the cron job, since I do it manually seldom enough that the proper flags aren't 
in my personal cache...


                --Steve Bellovin, http://www.cs.columbia.edu/~smb