Re: IOMEGA Rev intermittent failures

Brian K. White wrote:

The complexity of BE introduces failure points that I've hit that just
didn't even exist in ctar. In order to make my backups really reliable, I
had to make a script that knows how to find edge processes and kill them,
and find edge job lock files in the /usr/lib/edge tree and remove them, and
add that to cron just before the edge backup is scheduled. I never had to do
that with ctar and I was very incredulous to discover that a problem with a
bad tape or an ungracefully killed edge or some other transient problem one
night could cause _every subsequent backup to fail to even try to start_!
Due to a silly lockfile being left behind. To me that was way too delicate
of an arrangement for backup software. With ctar I would never have to worry
about that possibility of some customers box that is almost completely
un-monitored suffering a real crash and we discover that all 14 nightly
tapes were hopelessly obsolete because backups just stopped working one day
months or years ago because one day the server lost power while backups were
running. Every time the cron job fires up ctar, it's a completely fresh
attempt so no matter what happened yesterday, there is a as good a chance as
possible that today it will work. Transient problems are properly transient.
There are a few other things but I have no wish to slam BackupEdge. I merely
wish to show I'm not unduely biased in how I interpreted Tom's post.

These sound like serious issues with BackupEDGE, so I hope you have
properly communicated them to MicroLite. I hear a little detail bug
(with a big impact), a medium sized bug, and a serious behavior problem:

- not properly breaking stale lock files
- some sort of edge process hang (needs more detail)
- not adequately alerting the administrator when periodic scheduled
backups are failing

(From the way you describe it, I would guess these are problems you had
long ago with probably not the current version of BE. Which doesn't
mean they've necessarily been fixed since then -- but if not, again,
have you made sure ML knows about the problems?)

Bela<
.

Relevant Pages

  • Re: IOMEGA Rev intermittent failures
    ... didn't even exist in ctar. ... In order to make my backups really reliable, ... customers box where I first discovered it. ... I can't say that's the more proper behaviour for anyone else. ...
    (comp.unix.sco.misc)
  • Re: Backups failing on System State
    ... the past 4 backups have failed. ... it fails with that error. ... Media error is often a full media. ... repeatable for a transient problem like that. ...
    (microsoft.public.windows.server.sbs)