On Dec 3, 2006, at 3:44 PM, Tony Nelson wrote:
At 11:53 AM -0500 12/3/06, Jeff Johnson wrote:
Hint: attempting a cron script to "automate" --verifydb is going to
create more problems than fixes imho, judging from my experience
running rpm -qa from cron. That cron script is going to be ripped as
well,
Which script? The one on my site
<http://georgeanelson.com/rpm-verifydb.htm>, or a cron script that
is in
RPM now? Or am I misunderstanding what you mean by "ripped"?
AFAIK (not looked) you're thinking of doing rpm --verifydb from cron.
Well, rpm -qa has been done for years, and has always caused problems
because the cron script needs to open an rpmdb too. So any rpmdb
"hangs",
cron script runs for months, noone knows or cares.
Internalizing --verifydb, or externalizing rpm -qa output in /var/
cache/hrmib are
two approaches to not playing w Berkely DB and accomplishing the same
goals.
With fewer players, concurrent access in rpmdb is likelier to be
"stable" because
updates are like daily, not every second.
There's no known need to push the envelope like yum is doing to
handle ^C,
attempting to push concurrency to seconds instead of hours or days.
One does not have to fix problems that have not yet been seen, never
check
a return code that you don't know what to do with is a sane programming
paradigm.
73 de Jeff
_______________________________________________
Rpm-list mailing list
Rpm-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/rpm-list
