Developing filePro apps for developers
Brian K. White
brian at aljex.com
Sat Jun 16 11:19:48 PDT 2007
----- Original Message -----
From: "Fairlight" <fairlite at fairlite.com>
To: <filepro-list at lists.celestial.com>
Sent: Saturday, June 16, 2007 5:51 AM
Subject: Re: Developing filePro apps for developers
> I also didn't see any mention of it specifically, but I would -hope- that
> your period of quiescence is in effect for the -entire- update procedure,
> not just the restructure work. If you had processing in one place firing
> off processing in another place and there were interdependencies of
> action between the two and one processing table was updated but another
> was deferred till later in the update stream, then any actions taken in
> the interim could leave the customer with unintended data desyncs.
Thats just a detail among many.
Yes the procedure is a logical lock-out of the possibility of code running
while the system is in an inconsistent state.
And the system is now defined as the whole set of systems because we do now
have some things that transact from box to box.
so:
on all boxes, before doing any restructuring on the first or any box, do the
following in this order:
1) create /etc/nologin. now no new users can log in. use a message that
warns other root users what is going on.
2) shut down apache and cron, now no new cron jobs or cgi scripts can start.
edi files may be received by ftp still, but cron is not running so they will
just sit there harmlessly.
3) do 1 & 2 on all boxes to allow some time for already existing processes
to die naturally on boxes while you are going around beginning the lockdown
on the others.
4) back around to the first box, kill off any filepro-related processes or
processes that might start a filepro process. This just means kill all
clerks, reports, menus, start-scripts and login shells. I have a script to
do that easier than it sounds. do this on all boxes.
At this point no filepro processs will start on any box except by deliberate
interactive root user.
Nor will one of those rsync broadcasts go out by itself, but even if it did
they don't touch maps or indexes or keys or datas, so, they actually
shouldn't hurt even if they did.
5) Do your restructuring on all boxes.
6) Then restart the halted services and remove nologin on all boxes.
You don't reactivate any box untill all have been restructured.
It's possible to automate all of that but it will take a lot of care before
that script or system of cooperating scripts and network services is solid
enough to risk using for real.
Brian K. White brian at aljex.com http://www.myspace.com/KEYofR
+++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++.
filePro BBx Linux SCO FreeBSD #callahans Satriani Filk!
More information about the Filepro-list
mailing list