Last week, I requested information on mail spool backup strategies and I
received a large number of private responses. I want to express my
appreciation for all your feedback. We currently use a raided disk
topology and we just started doing twice daily incremental backups of our
mail spool directory and a full backup once a week. Eventually, we hope to
set up a separate system to do duty as a Networker backup server, but for
now, I think we are under reasonably good protection from any future
problems.
As far as recovering mail goes, we cannot stop our sendmail service for
the four hours or so that it would take to restore so much data. Doing so
would have caused far more problems (in the form of complaints from irate
users and inconvenience to those users). Hopefully, we will never lose our
mail spool directory again, but if we do, I will recover it to a different
subdirectory and than write a script to move the recovered files back into
place with sendmail off only for a few minutes while the script runs. The
idea of setting up different MX records for our system is interesting, but
I think it would complicate other matters due to the clustered system
environment in use here.
Thanks again
Received on Tue Mar 02 1999 - 21:48:28 NZDT