Veritas-bu

[Veritas-bu] Hard symbolic link restore issue...

2005-12-12 16:23:47
Subject: [Veritas-bu] Hard symbolic link restore issue...
From: Mark.Donaldson AT cexp DOT com (Mark.Donaldson AT cexp DOT com)
Date: Mon, 12 Dec 2005 14:23:47 -0700
There's an extensive discussion of the interaction of hard-links & Netbackup
in on pages 114-116 of the v5.x SAG (vol 1).

Hard-links, as you're probably aware, can't cross filesystem boundaries so
restoring to a mixed set of filesystems is bound to cause link issues.  In a
DR situation, I suspect your filesystems will be similar to your primary
setup and these problems might not manifest.

-M

-----Original Message-----
From: veritas-bu-admin AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu]On Behalf Of Christopher
Jay Manders
Sent: Monday, December 12, 2005 1:15 PM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] Hard symbolic link restore issue...



We have a restore problem that Legato (the old backup solution) did not 
seem to have with our iPlanet IMAP mail store ....

BACKGROUND:
We have a huge IMAP datastore that resides on a SAN. There are multiple 
front-ends for folks to get their mail from, named 'imap1, imap2,...'. 
When someone wants their mail spool files restored, those go onto a 
separate system that does not have all of the same data with the same 
inodes. So all Hard Links are not restoring properly

Here is a snippet of the error a standard restore gives:
09:41:59 (447272.005) /imap1-mailstore101/=user/c2/10/=+X+Xaa/98/69851.msg
09:41:59 (447272.005) Changed 
/imap1-mailstore101/=user/c2/10/=+X+Xaa/98/69851.msg to 
/store/restore/restore-XXaa//98/69851.msg
09:41:59 (447272.005) Could not link 
/store/restore/restore-XXaa/98/69851.msg -> 
/imap1-mailstore101/=user/85/c2/=+Y+Ybb/09/937.msg. Errno = 18: 
Cross-device link

QUESTION:
Is there a way to back up and restore the data in the links for these 
one-off requests for a user's mail store restore that we get daily, 
while also keeping the knowledge of the link..so in a REAL disaster we 
can restore to a system with the proper inode structures with the 
symbolic links intact?


TIA!


--Chris



_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

<Prev in Thread] Current Thread [Next in Thread>