Results 1 to 3 of 3

Thread: Outstanding Problems After a Week of Amanda

  1. #1
    Join Date
    Oct 2011
    Posts
    17

    Default Outstanding Problems After a Week of Amanda

    OK, so I've spent over the last week trying to get Amanda setup and working, and I'm still struggling to make sense of it all. If someone can help clarify a few things, that would be greatly appreciated.

    Environment: I have a dedicated amanda server doing the backups. It has an external DAS enclosure with 4tb of RAID. The server and all clients are running 3.3.0. The clients are a mix of Linux (CentOS and Fedora), Windows (2003, 2008 and XP) and SAMBA mounts (to a NAS device where I can't install the Amanda client).

    1. Holding disk is not being used.
    Despite having a dedicated mount (internal RAID SCSI drives) for the holding disk (which I read is still recommended, even with vtapes), it doesn't appear to get used; the logs don't mention it, amcheck doesn't mention it and nothing appears there while the backup is running.
    Code:
    /etc/amanda # cat DailySet1/advanced.conf 
    .....
    define holdingdisk caddy {
            comment "External Caddy RAID Array"
            directory "/amanda/holding"
            chunksize 256Mb
            use -64 Mb
    }
    .....
    
    
    /etc/amanda # df -h /amanda/holding/
    Filesystem            Size  Used Avail Use% Mounted on
    /dev/mapper/RAID0-amholding  116G  188M  110G   1% /amanda/holding
    
    
    /etc/amanda # cat template.d/dumptypes
    ......
    define dumptype user-tar {
        user-tar
        holdingdisk required
        compress client fast
    }
    ....
    You can see I have "holdingdisk required" in the dump definition. All entries in my disklist are using a dumptype that inherits the "user-tar" definition.
    Code:
    $ amcheck DailySet1
    Amanda Tape Server Host Check
    -----------------------------
    found in slot 5: volume 'DailySet1-5'
    slot 5: volume 'DailySet1-5' is still active and cannot be overwritten
    found in slot 6: volume 'DailySet1-6'
    slot 6: volume 'DailySet1-6'
    Will write to volume 'DailySet1-6' in slot 6.
    NOTE: skipping tape-writable test
    Server check took 2.315 seconds
    
    Amanda Backup Client Hosts Check
    --------------------------------
    Client check: 9 hosts checked in 2.855 seconds.  0 problems found.
    
    (brought to you by Amanda 3.3.0)
    2. Defining the "length" of a vtape.
    At the moment I have a tapetype definition in amanda.conf for HARDDISK which defines a "length" of the tape.
    Code:
    define tapetype HARDDISK {
           comment "Virtual Tapes"
           length 128 gbytes
    }
    What is the proper way to define the length as "use whatever the hell is free"? I could set the length to 4tb (the size of the DAS enclosure), or some stupidly huge value like 99999yottabytes but both of those seem like a hack solution.

    3. "Best practice" for defining a differential backup
    What is the generally accepted way to perform only full and differential backups? ie, only level 0 and level 1?

    4. Off-site Backup
    I have been able to find very little information on how to perform off-site backups (using more external hard drives). I think there are 2 possibilities:
    a) Use rsync to duplicate the vtape path and /etc/amanda/
    b) Use 'amvault' to duplicate a volume (vtape) to the external hard drive
    I want to end up with on-site, online copies of backups to restore deleted/corrupted files, and off-site, offline copies for disaster recovery.

  2. #2
    Join Date
    Mar 2010
    Posts
    27

    Default re: Holding disk not being used

    [url]http://forums.zmanda.com/showpost.php?p=9575&postcount=2[/url]

    ...has helpful details about holding disk, size and split params. Was part of thread a little different from your question, but thought it might help.

  3. #3
    Join Date
    Oct 2011
    Posts
    17

    Default

    After a bit of playing, it seems that the holding disk is being used, but nothing actually says that it is which is weird/annoying.

    With 'holdingdisk required', the backup dies with the following errors:
    driver: flush size 0
    find diskspace: not enough diskspace. Left with 7472000 K
    find diskspace: not enough diskspace. Left with 4314976 K
    find diskspace: not enough diskspace. Left with 414944 K
    find diskspace: not enough diskspace. Left with 2034816 K
    find diskspace: not enough diskspace. Left with 394816 K
    find diskspace: not enough diskspace. Left with 233792 K
    find diskspace: not enough diskspace. Left with 45920 K
    find diskspace: not enough diskspace. Left with 29856 K
    find diskspace: not enough diskspace. Left with 6944 K
    find diskspace: not enough diskspace. Left with 3744 K
    find diskspace: not enough diskspace. Left with 2464 K
    find diskspace: not enough diskspace. Left with 416 K
    find diskspace: not enough diskspace. Left with 160 K
    find diskspace: not enough diskspace. Left with 96 K
    find diskspace: not enough diskspace. Left with 96 K
    find diskspace: not enough diskspace. Left with 96 K
    find diskspace: not enough diskspace. Left with 96 K
    find diskspace: not enough diskspace. Left with 64 K
    find diskspace: not enough diskspace. Left with 64 K
    find diskspace: not enough diskspace. Left with 64 K
    find diskspace: not enough diskspace. Left with 64 K
    find diskspace: not enough diskspace. Left with 64 K
    find diskspace: not enough diskspace. Left with 64 K
    find diskspace: not enough diskspace. Left with 64 K
    find diskspace: not enough diskspace. Left with 64 K
    find diskspace: not enough diskspace. Left with 64 K
    Changing to 'holdingdisk auto' avoids these errors and allows the backup to run.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •