Backup Central

Dear BackupPC-developers and users,

I have two suggestions to improve a little the already
very good documentation of BackupPC.

First one: identical files
--------------------------

The term "identical files" is often used, of course,
because it is the core of this system. But nowhere in
the documentation it is said that "two identical files
are only identical in the sense of their contents,
independently of their attributes (srwx, ACL, chflags,
and so on)."

I think such a warning should be added to the
documentation (of course, I know that BackupPC takes
perfectly note of the attributes as lons as the
external, not BackupPC-dependent, transfer mechanism
is able to do it).

Second one: workflow
--------------------

BackupPC could be used in a "non-standard" mode, e.g.
say to cumulate old backup/archives already existing
(on tape or on disk). In this case the use is
independent of the time (and the retain and the number
of logs etc.) scheduling. Of course you can start it
again and again from the web interface.

But a clearly described workflow to do it manually (or
per batch files) should be mentioned in the
documentation.

I had to partially extract the required information to
do it from the perl scripts and to imagine the
necessary parameters (at least one perl script is not
up-to-day in his internal description: see
BackupPC_dump wich is taking a poorly described
parameter -f). Other workflow questions remains open
for me: e.g. should I run BackupPC_nightly after each
BackupPC_dump/BackupPC_link?

Perhaps has the author all these workflow questions
already answered for himself in constructing the
system (perhaps in graphical form).


What do you think about?


Erminio

---------------------------------------------------
Erminio BARANZINI, Muehlemattstr. 53, CH-3007 Berne,
Switzerland, aredat < at > yahoo.com
----------------------------------------------------



__________________________________________________
Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup
http://fifaworldcup.yahoo.com

_______________________________________________________________

Don't miss the 2002 Sprint PCS Application Developer's Conference
August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink

_______________________________________________
BackupPC-users mailing list
BackupPC-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/backuppc-users

I definately agree, but have been trying to be patient about it since
BackupPC is a relatively young project (ie. still lacks restores of
arbitrary directory trees via it's web interface so your lusers don't
have to bother you about them [actually, the command line restore tools
are a pretty sparse pain in the ass still, though better than nothing
when desperate]. Since I like the direction overall though, I'm living
with such things at present. But, yes, I could use a paragraph or two
on "design flow" myself.

-- Kyle

Thus Spake Erminio [Sun, 9 Jun 2002 04:24:39 -0700 (PDT)]

But a clearly described workflow to do it manually (or
per batch files) should be mentioned in the
documentation.

--
Kyle Amon email: amonk < at > gnutec.com
url: http://www.gnutec.com/~amonk
KeyID KeyID 1024D/4EB96E44
Fingerprint = E9EC 0046 8487 23D7 C91C D757 7B2A 8AE9 4EB9 6E44
"Proprietary and secret software is the moral equivalent of runners in
a fist fight. Sad to say, the only referee we've got does not seem to
object to fights; he just regulates them ("For every ten yards you run,
you can fire one shot"). He really ought to break them up, and penalize
runners for even trying to fight."

- Richard Stallman
The GNU Manifesto, 1985

This email Copyright 2001 by Kyle Amon, Inc., a Florida corporation,
6002 Palm Shadow Way, Suite 1218, Tampa, FL 33647: (813) 979-1311
amonk < at > gnutec.com http://www.gnutec.com/~amonk. All rights Reserved.

_______________________________________________________________

Don't miss the 2002 Sprint PCS Application Developer's Conference
August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink

_______________________________________________
BackupPC-users mailing list
BackupPC-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/backuppc-users

First one: identical files
--------------------------

The term "identical files" is often used, of course,
because it is the core of this system. But nowhere in
the documentation it is said that "two identical files
are only identical in the sense of their contents,
independently of their attributes (srwx, ACL, chflags,
and so on)."

I think such a warning should be added to the
documentation (of course, I know that BackupPC takes
perfectly note of the attributes as lons as the
external, not BackupPC-dependent, transfer mechanism
is able to do it).

Good idea. I'll add this.

Second one: workflow
--------------------

BackupPC could be used in a "non-standard" mode, e.g.
say to cumulate old backup/archives already existing
(on tape or on disk). In this case the use is
independent of the time (and the retain and the number
of logs etc.) scheduling. Of course you can start it
again and again from the web interface.

But a clearly described workflow to do it manually (or
per batch files) should be mentioned in the
documentation.

I had to partially extract the required information to
do it from the perl scripts and to imagine the
necessary parameters (at least one perl script is not
up-to-day in his internal description: see
BackupPC_dump wich is taking a poorly described
parameter -f). Other workflow questions remains open
for me: e.g. should I run BackupPC_nightly after each
BackupPC_dump/BackupPC_link?

First I would like to understand the goal here. As I understand it
you would like to have a scripting interface for starting backups.

I recommend against manually running BackupPC_dump/BackupPC_link
yourself, especially if BackupPC is running. There are several reasons
for this: BackupPC might run a backup on the same machine at the same
time you are manually backing it up by running BackupPC_dump. Also, at
most one BackupPC_link must run at any one time: the integrity of the
pool requires this. BackupPC (the main script, not the whole package)
enforces these constraints.

Perhaps the best way to manually run your own backups is to talk to
BackupPC and tell it what to do. That's what the CGI interface does.
This interface isn't documented, but it is pretty easy to look at the
CGI code to see what commands it sends to start backups.

In 1.5.0 the telnet method for connecting to BackupPC will no longer work.
There will be a new script, BackupPC_serverMesg, that takes a command-line
command and sends it to BackupPC. I will add some remarks in the docs
to describe how to use it to start a backup.

Does this meet your needs or do you have something bigger in mind?

Craig

_______________________________________________________________

Don't miss the 2002 Sprint PCS Application Developer's Conference
August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink

_______________________________________________
BackupPC-users mailing list
BackupPC-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/backuppc-users

Second one: workflow
--------------------

First I would like to understand the goal here. As
I understand it
you would like to have a scripting interface for
starting backups.

No, I am aiming only at what is already available.

BackupPC (the main script, not
the whole package)
enforces these constraints.

I am totally willing to run under BackupPC, no
problem. But I would like to do it from the shell and
not from the Web interface.

Does this meet your needs or do you have something
bigger in mind?

Craig

It is probably enough. I am not demanding to overbloat
the present well performing system.

Let me explain with an example. I have dozens of old
backups/archives tapes which are partially overlapping
and wich are identified by their date of creation. I
would gain profit from the very good pooling mechanism
of of BackupPC.

I imagine my workflow as follow (in form of a shell or
perl script):

# no normal backup is never running during following
operations
# the config file is adapted to this use
#
# start daemon
/usr/local/backuppc/bin/BackupPC -d
#
loop:
# read in tape content
cd /big_disk/oldies/tape_date; tar -x -f=/dev/nrsa0 .
# BackupPC
/usr/local/backuppc/bin/BackupPC_dump -f $host
/usr/local/backuppc/bin/BackupPC_link $host
/usr/local/backuppc/bin/BackupPC_nightly
/usr/local/backuppc/bin/BackupPC_trashClean
# remove old tape content
cd /big_disk/oldies/tape_date; rm -rf *
# rerun loop with next tape
goto loop

Of course, this is only in symbolic form. Tests of
identity of backup and origin and other sanity test
are necessary, warning to mount/unmount tapes are to
be issued and so on.

My questions:
- Could this be a workable skeleton?
- Am I missing steps in BackupPC?
- Are the BackupPC_nightly and BackupPC_trashClean
steps really needed for each call to the couple
BackupPC_dump/BackupPC_link?

Could you confirm that this is a safe method? And if
yes, could you think to put such a skeleton in the
documentation?

I thank you in advance

Erminio

---------------------------------------------------
Erminio BARANZINI, Muehlemattstr. 53, CH-3007 Berne,
Switzerland, aredat < at > yahoo.com
----------------------------------------------------



__________________________________________________
Do You Yahoo!?
Yahoo! - Official partner of 2002 FIFA World Cup
http://fifaworldcup.yahoo.com

_______________________________________________________________

Don't miss the 2002 Sprint PCS Application Developer's Conference
August 25-28 in Las Vegas - http://devcon.sprintpcs.com/adp/index.cfm?source=osdntextlink

_______________________________________________
BackupPC-users mailing list
BackupPC-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/backuppc-users