The thread is forked so I start a new Subject: line.
âChad,
There are other issues when you have a host fails in this manner/timing you
describe, but let's concentrate on the issues at hand, namely workfiles and
CSA database.
You don't really have to worry about your workfiles *unless* your CPs also
fail at the same time (in that case, call Schneider GCS and they'll help
you recover from that disaster: "Initialize. Reload!"). Once you restore
your host with your latest backup, just upload from your CP and the
workfile now matches what's running in the CP. Repeat for all CPs.
The Foxboro archive at freelists.org is full of tales of woes when old
workfiles get reloaded from old (ancient) backups. But your scenario
assumes that you have very recent backups so unless you've made extensive
edits all week, you shouldn't be in too bad of a shape when you reload from
your Sunday backup. (But do heed Alex Johnson's warning in this missive:
//www.freelists.org/post/foxboro/foxboro-CP-Image-WorkFile-and-CSA,3)
Joe Ricardi alluded to CSA_Save as a tool to back up your CSA database (in
/usr/fox/csa). Depending on how often you make backups and whether any
changes have been made to the CSA database since the last backup, you may
have to do some manual editing to get your database in sync, but it's not
as daunting as it sounds. CSA_Save and CSA_Merge are your friends, make
acquaintance with them before the time when you really need them so it's
not scary.
Duc
On Thu, Mar 23, 2017 at 9:49 AM, Chad Ziesemer <
cziesemer@xxxxxxxxxxxxxxxxxxx> wrote:
Jacques, excellent topic, thanks for starting the thread. This probably
only applies to ICC users: For those doing weekly back-ups, how do you
handle keeping the work-file, CSA, and CPs in sync if the box with the
work-file or the CSA fails and needs to be restored? I recently worked
through this and the procedures to get them synced back up was, well very
fun. (Aside: Huge thanks to all the content you've all added to the
freelist archives! It demoted a crisis to a learning experience). Say you
do this:
Back-up entire system on Sunday nights
Make CP edits on Monday - Thursday
Host server fails on Friday
Restore server to backup from Sunday
Your work file (and CSA if it's on the same box) is going to be missing
all of the edits from the past week.
In this case daily isn't even enough if there's an ill-timed failure. Is
this a case where incremental back-ups of these files should be used?
