We had a Solaris 8 system that had been up for 80 days.  We issued 5 "at -k
now" scripts in quick succession.  We found that cron would process only 1
script in the queue about every minute or so.

We then did "/etc/init.d/cron stop" followed by "/etc/init.d/cron start" and
the sluggishness went away; i.e., all 5 "at now" processes fired off within
a second.
Is this a know bug?

In article <GoIIFv....@wlbr.to.gd-es.com>,
        "Allen M. Cohen" <a...@to.gd-es.com> writes:

Don't know, but I recall "at now" being dicey for a _long_ time.  Why not
just use "batch" instead?  If you _must_ use "at" for some reason,
"at now + 1 minute" works well enough.

Also, see queuedefs(4) and consider changing the line for the "a" queue,
i.e. to something like

a.8j1n

which would allow more concurrent "at" jobs.

--
mailto:rlha...@mindwarp.smart.net  http://www.smart.net/~rlhamil

I did some research at sunsolve, and found that patch 109007-06 might fix
this.  I will let this group know if it works...
"Richard L. Hamilton" <rlha...@smart.net> wrote in message
news:u1t6phinf98j0a@corp.supernews.com...