The reason I was concerned about recyclable versus recoverable is
because we never recycle tapes that have these savesets on them. Once
they're backed up and cloned, we will never back them up again, and
there are no plans to recycle the tapes ever. So, at some point, the
saveset will be marked as recyclable, but it would be recoverable for a
year prior. Also, I have seen some cases where a saveset was flagged as
recyclable because of some premature abort or SCSI bus reset or some
such thing but could never really pin point what exactly caused it, but
I know it was only a few hours (days at the most) old.

In thinking about this a little more, though, it occurs to me that any
file that needs to be backed up by the script will be relatively recent;
so if even if the script determines that the file had been previously
backed up, it would be well within the last year, so it would be
recoverable and should not be recyclable. If it is in fact recyclable
then something's wrong, and it should be re-backed up anyway. The only
possible problem would be if someone created a new file with an older
name, say matching something that was backed up over a year ago. That
could be a problem because the script would see that as not recoverable,
but I guess you just need to make some reasonable assumptions and take
it from there.

George

Darren Dunham wrote:
>
> > Thanks, Darren. To be more specific, I don't care how long ago the
> > saveset was backed up. What I want to know is that if there is an entry
> > in the database for it, then I need to know if it's complete, not
> > suspect and either recoverable or recyclable. It's okay if it's
> > recylable as long as it's recylable because it's beyond the retention
> > policy (one year) and not for some other reason. If a saveset is
> > recylable, how do you know that is was not something other than time
> > that made it that way? Can NetWorker mark a saveset as recycable for
> > some other reason and if so, is there some attributes I can add to check
> > for that?
>
> I don't know any way that a saveset becomes recyclable other than
> passing the retention date (which may be modified after the fact), or by
> a manual process.
>
> Further, I don't think there's anything that says what caused it to
> become recyclable.
>
> > I guess I need to put some kind of 'or' condition on the command to
> > check for either recoverable or ssrecycle, but I don't see any way to do
> > that, so I'm thinking I would need to run the command first as:
>
> Well if you don't care if it's recyclable, then don't bother putting
> either one on the query line.
>
> > mminfo -s server -q
> > 'name=path,client=client,recoverable,!ssrecycle,!incomplete,!suspect' -r
> > 'client,ssid,savetime,name'
> >
> > and if it returns nothing then run it again as:
> >
> > mminfo -s server -q 'name=path,client=client,ssrecycle,!recoverable,
> > !incomplete,!suspect' -r 'client,ssid,savetime,name'
>
> I'd try... -q 'name=path,client=client,!incomplet,!suspect'
>
> > to be sure it's not still in there as recyclable and if this returns
> > something then check to see that it's older than one year? If not then
> > back up.
>
> I guess I don't understand why you care if one of them is over a year
> old but not the other.  Whether a 'recyclable' saveset is something you
> would want to worry about depends on how you deal with tapes, but I'd
> probably just look for recoverable (not recyclable) and backup if I
> didn't find it.
>
> That way when the recyclable saveset is finally recycled, I've already
> made my new copy and don't have to wait for it.
>
> > Looks like the mminfo command also supports the clflags showing aborted,
> > incomplete and suspect for clones savesets. Guess I can use that for
> > that.
>
> Note that every copy is a clone.  It's just that if you cloned a
> saveset, you'll have two different copies.  Some imagine that 'clflags'
> only reports on clone copies.
>
> You can view the flags on both the original and any subsequent copies.
> Make sure you can deal with both fixed and variable width fields.  You
> have to do 'clflags' as a fixed width field since it may have no output.
>
> Most of the other fields are easier to parse as variable-width.
>
> --
> Darren Dunham                                           ddunham AT taos DOT 
> com
> Unix System Administrator                    Taos - The SysAdmin Company
> Got some Dr Pepper?                           San Francisco, CA bay area
>          < This line left intentionally blank to confuse you. >
>
> --
> Note: To sign off this list, send a "signoff networker" command via email
> to listserv AT listmail.temple DOT edu or visit the list's Web site at
> http://listmail.temple.edu/archives/networker.html where you can
> also view and post messages to the list.
> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=