On Thu, Apr 28, 2005 at 10:06:15AM +0100, Chris Stoddart wrote:
>
> Hi gurus,
>
> I have a weird one I can't figure out. The situation is a Solaris
> machine with a small RAID attached. The RAID is partitioned into
> four 80Gb filesystems. The amanda server is also Solaris so we're
> using ufsdump to back them up.
Just a note, the OS of server plus client don't dictate backup
program, just client.
Funny what we call small these days. Of course I'm still in a SOHO.
> Now three of the partitions back up fine, but the fourth
> gives the [no backup size line] error consistently. The partition isn't
> fuller than the other three, the same binaries, tape drive, buffer disk
> are being used for all four, so why would just one partiton give this
> error?
>
> Below is the output from an amdump run. The only strange things I can see
> (apart from the error) is that it thinks that 75253.71MB needs to go on
> 1.12 tapes, when tapes are in fact 100Gb LTOs and the 'not enough space
> for slave process', whatever that means?
>
> /-- snarl /export/raid2 lev 0 FAILED [no backup size line]
> sendbackup: start [snarl:/export/raid2 level 0]
> sendbackup: info BACKUP=/usr/local/etc/amufsdump
> sendbackup: info RECOVER_CMD=/usr/sbin/ufsrestore -f... -
> sendbackup: info end
> | DUMP: Writing 32 Kilobyte records
> | DUMP: Date of this level 0 dump: Wed Apr 27 18:31:19 2005
> | DUMP: Date of last level 0 dump: the epoch
> | DUMP: Dumping /dev/rfssnap/1 to standard output.
> | DUMP: Mapping (Pass I) [regular files]
> | DUMP: Mapping (Pass II) [directories]
> | DUMP: Estimated 154119604 blocks (75253.71MB) on 1.12 tapes.
> | DUMP: Cannot create slave process: Not enough space
> | DUMP: The ENTIRE dump is aborted.
> sendbackup: error [no backup size line]
> \--------
>
> If anyone has any ideas what I could be looking for, I'd be most grateful.
>
I don't recognize the messages. So just directions to suggest.
The messages are coming from ufsdump, not amanda.
It is ufsdump failing, then of course amanda fails.
I'd concentrate on that aspect, why ufsdump fails.
The actual command line is in the debug files (likely /tmp/amanda)
on the client. Maybe additional info there, and you can try to
run the actual command by hand.
An 80GB filesystem that is filled to 75.25GB is pretty full.
Sometimes filesystems have problems when close to full.
I see you are using filesystem snapshots. Is your ?backing store?
of sufficient size? Might that have anything to do with the
"slave process"?
Amanda knows the size of your tape device. Does ufsdump?
The 1.12 tapes message is coming from ufsdump, not amanda.
You are using a ufsdump wrapper, amufsdump. Might there be
any oddities in the wrapper?
It appears this is the first dump of the FS.
Is the problem repeatable?
--
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)
|
