This is something that Legato needs to document more completely. We should
not need to run tests to understand how Legato expires and recycles
savesets, etc.
Legato Staff:
If you have an internal document that treats such matters and it can be
released, would you pleas make it available to this group. If the full
document cannot be released, would someone cut and paste the appropriate
sections from the document that treats these issues.
-----Original Message-----
From: George Sinclair [mailto:George.Sinclair AT NOAA DOT GOV]
Sent: Friday, May 30, 2003 4:31 PM
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Subject: Re: [Networker] Why is an old saveset still browsable?
Yes, this appears to be exactly what's happening. It looks as if the
problem occurs in the following two scenarios:
1. The client is no longer being backed up, never mind the fact that its
index may or may not still exist on disk.
2. The affected saveset is no longer being backed up -- either it
doesn't exist anymore or it's just being skipped -- but the client is
still being backed up.
I ran an mminfo command to track these rascals down. Excluding cases
where the saveset is still within the browse policy, I still found a ton
of them. It looks like most of the "recoverable" ones show up with "vrF"
or sometimes "vrEiF" for the ssflags whereas these ones that are still
listed as "browsable" seem to show up with "vF". So I greped out the
ones with "vrF" and "vrEiF" and the majority of the output fit into one
of the above two scenarios. The rest were divided between those that
were still legitimately within the browse policy and those that were
actually "recoverable" but showed up with "vF". In most cases, however,
this was due to the fact that they were written to a pool that had
indexing turned off so the client's index did not get updated. Not sure
about the others.
George
Davina Treiber wrote:
>
> On Wed, 28 May 2003 18:35:04 -0400, George Sinclair
> <George.Sinclair AT NOAA DOT GOV> wrote:
>
> >I have some savesets that are from 10/??/02 but still show up as
> >browsable on several older volumes. I can see this from the volumes
> >listing under nwadmin. This seems odd given the fact that the browse
> >policy for the affected clients was one month. One of the clients is no
> >longer a member of any active group, but the index has not been removed.
> >The other client is no longer listed in the client listing under
> >nwadmin, and it's client index is gone. I guess I could understand this
> >occurring for the one index that still exists because perhaps NetWorker
> >still sees the last runs as browsable since the index is no longer
> >getting updated, so:
> >
> >1. Am I wrong about that?
> >
> >2. Why would some of the savesets for the other index that no longer
> >exists still be identified as browsable?
>
> Basically, it's a bug. nsrim does not work on save sets that no longer
have
> a client resource associated with them. Legato know about this problem and
> I expect the fix has gone into some reecent versions. I had a hot fix for
> certain versions, but if there's only one or two clients giving you this
> problem it's probably easier for you just to run nsrim for the problem
> client, using: nsrim -c client. This should do the trick.
>
> --
> 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.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--
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.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
|
