if amfetchfump will not disturb a concurrent taper, I can fix our
documentation to use it instead. (and indeed, I can run multiple
amfetchdumps simultaneously, using different tapes)

Thank you


On Mon, May 5, 2014 at 1:04 PM, Jean-Louis Martineau
<[email protected]>wrote:

>
> You can also try this patch which allow to read a SPLIT_DUMPFILE from
> holding disk.
>
> Jean-Louis
>
>
> On 05/05/2014 12:41 PM, Jean-Louis Martineau wrote:
>
>> amrestore is designed to work with holding disk file or device.
>> You use it with a vtape file, amrestore think it is an holding file and
>> to validation on it.
>>
>> In 3.3.5, amrecover, amfetchdump and amrecover should works even if a
>> backup is in progress, you should use:
>>   amfetchdump CONF xxx.cmu.edu imap/lvol0 20140504200001
>>   amrestore chg-disk:/amanda/vtapes xxx.cmu.edu imap/lvol0 20140504200001
>>
>> Jean-Louis
>>
>> On 05/05/2014 11:55 AM, Chaskiel Grundman wrote:
>>
>>> amrestore -p /amanda/vtapes/slot1/00001.xxx.cmu.edu.imap_lvol0.0 >
>>> /dev/null
>>> Reading from '/amanda/vtapes/slot1/00001.xxx.cmu.edu.imap_lvol0.0'
>>> split dumpfile: date 20140501110923 host xxx.cmu.edu <http://xxx.cmu.edu>
>>> disk imap/lvol0 part 1/UNKNOWN lev 0 comp N program /sbin/dump
>>> ERROR: unexpected header type 5 in holding file
>>> '/amanda/vtapes/slot1/00001.xxx.cmu.edu.imap_lvol0.0'
>>>
>>>
>>>
>>>
>>> On Mon, May 5, 2014 at 11:44 AM, Jean-Louis Martineau <
>>> [email protected] <mailto:[email protected]>> wrote:
>>>
>>>     Chaskiel,
>>>
>>>     Can you post the exact amrestore command line and the exact error
>>>     you get.
>>>
>>>     Jean-Louis
>>>
>>>
>>>     On 05/05/2014 10:43 AM, Chaskiel Grundman wrote:
>>>
>>>         All lines were
>>>
>>>                 ALLOW-SPLIT         no
>>>
>>>         and the tapefile headers are still split:
>>>         AMANDA: SPLIT_FILE 20140504200001 xxx.cmu.edu
>>>         <http://xxx.cmu.edu> <http://xxx.cmu.edu> imap/lvol0  part
>>>         1/-1  lev 0 comp N program /sbin/dump
>>>
>>>
>>>         I am having trouble figuring out the perl->C interface. I
>>>         think the problem might just be that taper always fills in a
>>>         part number in the dump_header, but I haven't found where the
>>>         perl dump_header hash is turned into a dumpfile_t
>>>
>>>
>>>         On Fri, May 2, 2014 at 1:35 PM, Jean-Louis Martineau
>>>         <[email protected] <mailto:[email protected]>
>>>         <mailto:[email protected] <mailto:[email protected]>>>
>>>         wrote:
>>>
>>>             On 05/02/2014 11:11 AM, Chaskiel Grundman wrote:
>>>
>>>                 On our old installation (2.6.0p2) , we used amrestore -p
>>>                 <vtape file>, not amfetchdump or amrecover, to
>>>                 decode/decompress dumps when doing restores. This
>>>         works even
>>>                 when a backup job is in progress.
>>>
>>>                 This is not working on 3.3.5, because the vtape files
>>>         are all
>>>                 split dumps (with one part), and amrestore does not
>>>         understand
>>>                 them. Is there any way to force taper to not use the
>>>         splitter?
>>>                 My configuration has runtapes 1, and I am using chg-disk.
>>>                 Setting the dumptype to 'allow-split no' had no effect.
>>>
>>>
>>>             Are you sure it is correctly set?
>>>               run: amadmin CONF disklist | grep -i allow-split
>>>
>>>             Jean-Louis
>>>
>>>
>>>
>>>
>>>
>>
>

Reply via email to