[rescue] Mounting and Dumping
Sheldon T. Hall
shel at cmhcsys.com
Wed Jan 14 13:04:39 CST 2004
Janet L. Campbell writes ...
> On Tue, 13 Jan 2004, Sheldon T. Hall wrote:
> > So ... in reading the ufsdump doco for Solaris 7, it really
> > emphasizes the
> > need to dump only UNmounted filesystems. So my questions are ...
> > 1) Is that really neccessary?
> It depends on what's happening to the filesystems at the time. The
> principle danger of dumping a mounted filesystem is that writes may make
> the dump inconsistent. This is *usually* not a big problem when the
> machine is otherwise quiet.
Inconsistent, I can stand, Unreadable, I can't. Will the fs activity while
dumping a mounted filesystem make the dump unusable, or just make a couple
of the files inconsistent? What about the tape's directory structure? Any
chance of something there that will muck things up?
> > 2) If so, can I unmount, dump, and re-mount filesystems under
> > script/batch control?
> If you could interactively, sure. It's probably overkill, though.
> > 3) All of 'em? If I unmount /usr or /, how do I run the program?
> With difficulty. It's really not necessary to quiesce write access as
> long as you don't really care about the writing that is in progress during
> the dump.
> > 4) What's a fellow to do?
> Well, fssnap is quite nifty. It solves this problem nicely.
When I finally upgrade to Solaris 9, I'll use it. fssnap doesn't seem to be
available for Solaris 7.
> > I have separate filesystems for /, /usr, /var, /users (everyone's home
> > directory), etc., but they are all on one disk. In running a test, it
> > sounded like the Classic wasn't keeping up with the DLT drive. Would
> > putting the DLT on the the SunSwift card's SCSI bus help that?
> Maybe. If you're dumping a partition with lots of small files, the
> Classic may just not be able to keep up with the DLT drive because of
> disk speed and fragmentation.
Jochen Kunz suggests the same thing: the Classic isn't as fast at the tape.
He also suggested disableing compression to help it stream, and I'll try
Clayton Wheeler suggests using "iostat -ncx 10" to show where the bottleneck
is, and I'll probably do that, as well.
> If you have a bunch of scratch space, you can use this to "stage"
> dumps by dumping to this space and then dd'ing off to tape. Some
> backup suites, such as Amanda, do something similar.
I don't currently have space on that machine to "stage" the dump, but I've
thought of adding another disk for that purpose. That would also give me a
much faster way to restore recent data, which would have benefits beyond the
More information about the rescue