RE: increment only config question

[Mark Holm]

Did anybody ever give you an answer to this as it is the same type of setup
that I am attempting to do as well? The book mentions that you need to do
something with amadmin -force, but I thought that was how you forced it to
DO a full dump.

markh

-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
Sent: Friday, September 07, 2001 1:57 AM
To: [EMAIL PROTECTED]
Subject: increment only config question


I understand how Amanda will spread full and incremental backups along the
dumpcycle depending on dumpcycle, tapecycle, and runspercycle values.
However, let us assume that we want Amanda to perform full backups on
Sundays and incrementals on the remaining days.

I can use a separate "Full" config that performs dumpcycle 0's whenever it
is run. My question is that for the incrementals,even if I have the
following below as a separate dumptype under the "Weekdays" config,


define Weekdays { 
program "GNUTAR"
compress none
index
strategy incronly   
} 

"Weekdays" config will always have to initally make a level 0 dump since it
can not see the index info generated by the "Full" config, all subsequent
backups after this will be incremental. There is no way I can avoid making
the extra initial full backup by allowing the "Weekdays" config to see the
index created by "Full" config right?


Thanks.

end

how to unsubscribe

[Bernhard R. Erdmann]

I just spent the whole day upgrading/hacking Cyrus for inserting proper
Return-Path:-headers just to make Sieve's reject work...

Part of my ~/.sieve now contains

elsif allof (address :contains "to"
["[EMAIL PROTECTED]","[EMAIL PROTECTED]"], header :is
"subject" "unsubscribe") {
reject "please visit http://www.amanda.org/ for instructions how to
unsubscribe from [EMAIL PROTECTED]";
}

Hopefully people will get to know how to unsubscribe.

No index records for host on date xxxxx

[Dan Smith]

OK, the server that runs the amanda stuff is what gets backed up.  I
added another host to the list, flare:/etc.  It seems to backup fine;
I get a size and time and everything in the email reports.  When I run
amrecover, I can sethost and setdisk, but there is no index records
for the host on any backup date.  The other machine is solaris 8.  The
amanda box is Linux.  If I look in daily/index/flare/_etc, there are
.gz files for all the dates.

Can someone help me diagnose this?

--Dan

Re: Configuring an Exabute 10h

[Thomas Hepper]

Hi,
On Fri, Sep 07, 2001 at 11:53:45AM -0700, Jeremy Wadsack wrote:
> 
> Thanks for the steps, the results are getting closer.

No Problem :-)

> > 3. Try chg-scsi -scanbus (i'm not sure that it works :-)
> 
> This doesn't work on my chg-scsi (2.4.2p2).

OK, so this is only in the current CVS version.

> 
> Ok, my changer is at /dev/sg0 (and tape at /dev/sg1). You'd think that
> the device numbering would follow chains, ids, etc. but I guess not.

To  be sure you can check your syslog files to see which id is assigned
to which device from the sg driver

> 
> 
> 
> > 6. After this try an amtape daily reset
> 
> Returned
> 
> amtape: changer is reset, slot chg-scsi: is loaded.

Ups, i think this should look like
[root@skywalker /root]# amtape-2.4.2p2 exabyte reset
amtape: changer is reset, slot 1 is loaded.

> 
> 
> So it looks like this might work. How does this approach compare to
> Paul's chg-userland fix?

chg-scsi does all the error handling and talk with SCSI commands with the
devices. It does not need special modules in the kernel which does the error
handling etc. In chg-scsi you can very easy define new driver for 
tape libs if they use non standard ways for moving or error handling. 
(Barcode reader for example)


Thomas

-- 
  ---
  |  Thomas Hepper[EMAIL PROTECTED] |
  | ( If the above address fail try   ) |
  | ( [EMAIL PROTECTED])|
  ---