[Rpm-maint] Removing repackage and (auto)rollback?
James Olin Oden
james.oden at gmail.com
Wed Feb 20 16:20:11 UTC 2008
On Feb 19, 2008 8:54 AM, James Antill <james.antill at redhat.com> wrote:
> On Tue, 2008-02-19 at 11:15 -0500, seth vidal wrote:
> > Do you like the perform an lvm-snapshot in pre-trans and allow rolling
> > back in post-trans as reliable rollback method?
> The problem is that while FS LVM snapshot+rollback might be reliable
> from the POV of rpm, it's untenable from the POV of anything else that's
> running on the system at the same time.
> Which basically makes it worthless, IMO.
Please explain your reasoning. I don't get why its untenable.
BTW, I have actually used rpm rollbacks in the field and they
basically work, because one of a few reasons:
- Most scriptlets in RedHat packages are fairly simple so their
typically is nothing to really undo.
- In house scriptlets were always written with the view that they
were going to be in rolled back so we made them work.
- Often we could work around an upstream scriptlets issuses in
rollback via some other rpm, or some code external to rpm.
That said rpm rollbacks never were turn-key in that you really had to
pay attention to many many details in order to make sure that when the
time came to turn the crank and turn time back, it would go off
without a hitch. Honestly, the current implementation isn't that bad
(especially in the rpm5 code base, as I spent much time in the
pre-rpm5 code making things better). If there was someone really
dedicated to it (i.e. for a period of about 6 months to a year, it
could be made even better.
That said it will never be as reliable and simple as a rollback of a
filesystem snapshot (however that is implemented). OTOH, that solution
is outside the rpm problem space.
> James Antill <james.antill at redhat.com>
> Red Hat
> Rpm-maint mailing list
> Rpm-maint at lists.rpm.org
More information about the Rpm-maint