Amanda-Users

Re: Trying to backup to disk, getting "dump to tape failed, will try again"

2007-11-27 11:06:57
Subject: Re: Trying to backup to disk, getting "dump to tape failed, will try again"
From: Jon LaBadie <jon AT jgcomp DOT com>
To: amanda-users AT amanda DOT org
Date: Tue, 27 Nov 2007 10:59:20 -0500
On Tue, Nov 27, 2007 at 09:42:30AM -0500, Jennifer Luisi wrote:
> Hi there,
> 
> I'm new to Amanda and am trying to get a disk-to-disk backup going.  
> When I run amcheck, everything seems fine.  Then, when I run amdump, 
> nothing ever gets written to disk (virtual tape).  The amanda log file 
> shows:
> 
> SUCCESS taper [hostname] /export/backup 20071126 0 [sec 188.985 kb 0 kps 
> 0.0 {wr: writers 0 rdwait 188.984 wrwait 0.000 filemark 0.000}]
> INFO driver [hostname] /export/backup 0 [dump to tape failed, will try 
> again]
> 
> I've googled and read docs and how-to's to get where I am now.  
> Unfortunately, all the docs and how-to's get me to running amcheck 
> without errors and then launch into how to start a recovery process.  I 
> guess that means that most people have no problems with amdump. ;-)  
> I've included config info below, amanda.conf, disklist, etc.  Please 
> take a look and let me know what silly error I've made.  I'd really 
> appreciate it.  If I've left out some pertinent info, please tell me and 
> I'll include it.
> 
> Thanks!!

Not sure that any of my comments will fix your problem,
but I just include a few notes.


> 
> My amanda.conf file is
> org "OrgName" # your organization name for reports
> dumpuser "amanda" # the user to run dumps under
> 
> inparallel 4 # maximum dumpers that will run in parallel
> 
> dumpcycle 14 # the number of days in the normal dump cycle
> runspercycle 10 # the number of amdump runs in dumpcycle days
> etimeout 55 # number of seconds per filesystem for estimates.

55 seconds to do a estimate is VERY short.  Some who use tar
and have it traverse the file system need hours (N * 3600)

> runtapes 1 # number of tapes to be used in a single run of amdump
> tapetype HARDDISK # what kind of tape it is (see tapetypes below)
> labelstr "^OrgName-[0-9][0-9]*$" # label constraint regex: all tapes 
> must match
> 
> tapecycle 11

That is a bare minimum.  You will essentially have no redundancy.
And if things mess up two days in a row, you will have lost some
of your only full dumps.

> tpchanger "chg-disk"
> changerfile "/export/amanda/status/chg-disk-status" # status files prefix
> tapedev "file:/export/amanda/vtapes/tape1/slots"
> 
> infofile "/export/amanda/adm/BlueFrog/curinfo"  # database filename
> logdir   "/export/amanda/adm/BlueFrog"          # log directory
> indexdir "/export/amanda/adm/BlueFrog/index"    # index directory
> 
> define tapetype HARDDISK {
>    comment "Dump onto hard disk"
>    length 40960 mbytes
> }

11 tapes by 40 GB, I presume this is a 500GB disk?
Maybe external?

> 
> define dumptype global {
>    index yes
>    program "GNUTAR"
>    exclude list "/var/lib/amanda/exclude.gtar"
>    # record no

If you don't allow recording, you can't use amrecover
to retrieve individual files convienently.

Whoops, that indexing, ignore above comment.

If you don't record your dumps, amanda will always
think you need a level 0 dump.  It will not have a
record of previous dump levels.


> }
> 
> define dumptype hard-disk-dump {
>    global
>    comment "Back up to hard disk instead of tape using dump"
>    holdingdisk no

A holding disk is still useful to speed things up by allowing
multiple dumps in parallel to the holding disk.  Without a
hold disk, you can only dump one DLE at a time and that may
be speed constrained by the network or other factors.

>    index yes
>    priority high
>    kencrypt no
>    program "GNUTAR"
> }
> 
> define interface vbeb {
>    comment "10 Mbps ethernet over the vbeb"
>    use 800 kbps
> }
> 
> 
> And my test disklist file looks like this:
> client.host /etc hard-disk-dump -1 vbeb
> client.host /export/backup hard-disk-dump -1 vbeb
> client.host /export/subversion hard-disk-dump -1 vbeb
> client.host /export/httpd/beta-webspaces hard-disk-dump -1 vbeb
> client.host /export/httpd/swc.bluefrog hard-disk-dump -1 vbeb
> 
> I have my virtual tapes set up and labeled:
> [root@backuphost ~]# ls -lR /export/amanda/vtapes/ | more
> /export/amanda/vtapes/:
> total 4
> drwxr-x---  4 amanda disk 4096 Nov 26 10:41 tape1
> 
> /export/amanda/vtapes/tape1:
> total 12
> drwxr-xr-x   2 amanda disk 4096 Nov 23 13:19 data
> -rw-------   1 amanda disk   11 Nov 23 13:19 info
> drwxr-xr-x  12 amanda disk 4096 Nov 26 13:53 slots
> 
> /export/amanda/vtapes/tape1/data:
> total 0
> 
> /export/amanda/vtapes/tape1/slots:
> total 44
> lrwxrwxrwx  1 amanda disk   39 Nov 26 13:53 data -> 
> /export/amanda/vtapes/tape1/slots/slot4
> -rw-------  1 amanda disk   11 Nov 26 13:53 info
> drwxr-xr-x  2 amanda disk 4096 Nov 26 12:38 slot1
> drwxr-xr-x  2 amanda disk 4096 Nov 26 10:51 slot10
> drwxr-xr-x  2 amanda disk 4096 Nov 26 13:29 slot2
> drwxr-xr-x  2 amanda disk 4096 Nov 26 13:51 slot3
> drwxr-xr-x  2 amanda disk 4096 Nov 26 15:12 slot4
> drwxr-xr-x  2 amanda disk 4096 Nov 26 10:49 slot5
> drwxr-xr-x  2 amanda disk 4096 Nov 26 10:51 slot6
> drwxr-xr-x  2 amanda disk 4096 Nov 26 10:51 slot7
> drwxr-xr-x  2 amanda disk 4096 Nov 26 10:51 slot8
> drwxr-xr-x  2 amanda disk 4096 Nov 26 11:40 slot9
> 
> 

Can you manipulate your vtapes with various sub-commands
of amtape?


-- 
Jon H. LaBadie                  jon AT jgcomp DOT com
 JG Computing
 4455 Province Line Road        (609) 252-0159
 Princeton, NJ  08540-4322      (609) 683-7220 (fax)