Hi,
The file: driver will not work with the rait: driver, the problem
is that the rait fork many process to do command, but many of the file:
command change global variable, these change are lost when executed
under a forked process.
Jean-Louis
On Wed, Feb 19, 2003 at 11:46:51AM -0500, Milos Brkic wrote:
> Hi,
> I sent this a while ago, and have gotten no responses. Does that mean
> that nobody has really done mirroring to disk and tape successfully, and
> I should move on to finding a different solution? What I am trying to do
> is keep the latest set of bakcups on disk, while sending the tape copy
> offsite. I have seen a suggestion to make hardlinks to the backups on
> holding disk before amflush is executed. Does anyone have the script
> changes required to do this already made, so I don't have to rediscover
> the wheel?
>
> Cheers,
> Milos
>
> The following is a repost of a description of what happens when I tried
> to set up disk and tape mirrors of backups:
>
> ...
>
> Here are some details on what I have done, and the behaviour that I
> get. If you need more info/details, please let me know the
> specifics. Otherwise, can someone tell me what I am missing?
>
> In the amanda.conf:
> ===========================================================
> ...
> runtapes 1 # number of tapes to be used in a single run
> of amdump
> tapedev "rait:{file:/backups/current1,tape:/dev/rmt/0bn}" #
> mirrored backups
> rawtapedev "/dev/null" # the raw device to be used (ftape only)
>
> tapetype BIGDAT # what kind of tape it is (see tapetypes below)
> labelstr "^flxmth[0-9][0-9]*$" # label constraint regex: all tapes
> must match
> ...
> define tapetype BIGDAT {
> comment "BIGDAT tape drives"
> length 20000 mbytes # these numbers are not accurate
> filemark 100 kbytes # but you get the idea
> speed 100 kbytes
> }
> ...
>
> The rest of the config should be standard/irrelevant.
> =============================================================
> Here is what happens when I try to run amlabel:
>
> amanda@matrix $ pwd
> /backups/current1
> amanda@matrix $ ls -la
> total 18
> drwxr-xr-x 7 amanda amandagrp 512 Jan 30 14:58 .
> drwxr-xr-x 12 root sysadmin 512 Jan 27 17:38 ..
> -rw-r--r-- 1 amanda amandagrp 0 Jan 30 14:58 am.log
> lrwxrwxrwx 1 amanda amandagrp 9 Jan 30 14:58 data -> flxmth005
> drwxr-xr-x 2 amanda amandagrp 512 Jan 28 09:43 flxmth001
> drwxr-xr-x 2 amanda amandagrp 512 Jan 28 12:11 flxmth002
> drwxr-xr-x 2 amanda amandagrp 512 Jan 30 12:39 flxmth003
> drwxr-xr-x 2 amanda amandagrp 512 Jan 30 14:48 flxmth004
> drwxr-xr-x 2 amanda amandagrp 512 Jan 30 14:58 flxmth005
> -rw------- 1 amanda amandagrp 11 Jan 30 14:53 info
> amanda@matrix $ ls -l flxmth005
> total 0
> amanda@matrix $ more info
> position 1
> amanda@matrix $ ammt -t file:/backups/current1 rewind
> amanda@matrix $ mt -t /dev/rmt/0bn stat
> Vendor 'SEAGATE ' Product 'DAT 9SP40-00' tape drive:
> sense key(0x0)= No Additional Sense residual= 0 retries= 0
> file no= 0 block no= 64
> amanda@matrix $ mt -t /dev/rmt/0bn rewind
> amanda@matrix $ amlabel -f flxmth flxmth005
> rewinding, reading label, reading label: Error 0
> rewinding, writing label flxmth005, checking label
> amlabel: reading label: Error 0
> amanda@matrix $ mt -t /dev/rmt/0bn stat
> Vendor 'SEAGATE ' Product 'DAT 9SP40-00' tape drive:
> sense key(0x0)= No Additional Sense residual= 0 retries= 0
> file no= 0 block no= 64
> amanda@matrix $ ls -l flxmth005
> total 132
> -rw------- 1 amanda amandagrp 10 Jan 30 15:01 00000-unknown
> -rw------- 1 amanda amandagrp 32768 Jan 30 15:01 00000.unknown
> -rw------- 1 amanda amandagrp 10 Jan 30 15:01 00001-unknown
> -rw------- 1 amanda amandagrp 32768 Jan 30 15:01 00001.unknown
> amanda@matrix $ more info
> position 2
> position 2
> ====================================================================
> This already seems wrong. Here is what happens when you then try to
> run an amdump:
>
> amanda@matrix $ mt -t /dev/rmt/0bn rewind
> amanda@matrix $ ammt -t file:/backups/current1 rewind
> amanda@matrix $ amdump flxmth doom c0t0d0s3
> amanda@matrix $ ls -l flxmth005
> total 70340
> -rw------- 1 amanda amandagrp 10 Jan 30 15:01 00000-unknown
> -rw------- 1 amanda amandagrp 32768 Jan 30 15:01 00000.unknown
> -rw------- 1 amanda amandagrp 13 Jan 30 15:04 00001-unknown
> -rw------- 1 amanda amandagrp35946496 Jan 30 15:04 00001.unknown
> amanda@matrix $ mt -t /dev/rmt/0bn stat
> Vendor 'SEAGATE ' Product 'DAT 9SP40-00' tape drive:
> sense key(0x0)= No Additional Sense residual= 0 retries= 0
> file no= 3 block no= 0
> =====================================================================
>
>
> Cheers,
> Milos
>
>
--
Jean-Louis Martineau email: martineau AT IRO.UMontreal DOT CA
Departement IRO, Universite de Montreal
C.P. 6128, Succ. CENTRE-VILLE Tel: (514) 343-6111 ext. 3529
Montreal, Canada, H3C 3J7 Fax: (514) 343-5834
|