On Sunday 16 November 2008, Dustin J. Mitchell wrote:
>On Sun, Nov 16, 2008 at 7:05 AM, Gene Heskett <gene.heskett AT verizon DOT
>net>
wrote:
>> I rebooted to 2.6.27.6 yesterday, from 2.6.28-rc4, and although the ls -l
>> of /dev/sd* does not change, amanda thought the whole system needed a
>> level 0 disguised as a level 1.
>
>Recall that it's not the device numbers of the /dev/sd* that will
>change -- it's the devices from which you're mounting your
>filesystems, which IIRC are LVM volumes. So you should be checking
>the major/minor numbers of those device files.
>
Not using LVM here, turned out to be a bigger headache than it solved for me.
>If the --no-check-device option is not working correctly, then it
>would best be demonstrated with a small example using *only* tar, and
>leaving Amanda out of the situation.
From this mornings /tmp/amanda-dbg/1226818737.389798: amgtar: pid 23721 ruid
501 euid 501 version 2.6.1alpha: start at Sun Nov 16 01:58:57 2008
1226818737.389856: amgtar: version 2.6.1alpha-20081114
1226818737.389978: amgtar: GNUTAR-PATH /bin/tar
1226818737.389984: amgtar: GNUTAR-LISTDIR /usr/local/var/amanda/gnutar-lists
1226818737.389995: amgtar: ONE-FILE-SYSTEM yes
1226818737.389999: amgtar: SPARSE yes
1226818737.390003: amgtar: ATIME-PRESERVE yes
1226818737.390010: amgtar: CHECK-DEVICE yes
1226818737.390016: amgtar: SIZE ^ *Total bytes written: [0-9][0-9]*
1226818737.390021: amgtar: IGNORE : Directory is new$
1226818737.390024: amgtar: IGNORE : Directory has been renamed
1226818737.390028: amgtar: NORMAL ^could not open conf file
1226818737.390032: amgtar: NORMAL ^Elapsed time:
1226818737.390036: amgtar: NORMAL ^Throughput
1226818737.390039: amgtar: NORMAL : socket ignored$
1226818737.390043: amgtar: NORMAL : File .* shrunk by [0-9][0-9]* bytes,
padding with zeros
1226818737.390052: amgtar: NORMAL : Cannot add file .*: No such file or
directory$
1226818737.390056: amgtar: NORMAL : Error exit delayed from previous errors
Note CHECK-DEVICE yes...
However, the corresponding runtar-*-debug files contain no reference
to --no-check-device. Oh, wait, those stop at Nov 13th, ignore.
Last nights selfcheck says 'check-device "no" in pipeline'
As does sendbackup, on 2 separate lines.
Also sendsize says "no"
Now I am totally bumfuzzled. Unless that option is not available in my
locally built tar-1.20. In which case should it not fuss about a bad option?
Before I was just confused.
>I think you determined that the holding-disk problems were a disk failure.
And I have badblocks running on it now, which is loading up the messages log,
but so far has a zero length badblocks-sda file on a usbkey. It is a little
over half done. And, for all the errors in the messages logfile while
badblocks is running, smartctl --all /dev/sda is not showing an increment in
the error count it is displaying. Call me puzzled, but I probably won't
answer. At least not till badblocks either upchucks and dies, or finishes
and writes a list I can feed e2fsck to fix. But in that event, I already
shined myself up to go to town and get another drive. Working in the
shop/garage I built last summer during the warmer parts of the day, trying to
convert a pile of cherry into an entertainment center is taking what little
energy I have left at 74 and counting. That and this machine "keeps me out
of the bars" :)
See <http://gene.homelinux.net:85/gene/E-Center> for a few pix. :)
Thanks Dustin.
--
Cheers, Gene
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Under capitalism, man exploits man. Under communism, it's just the opposite.
-- J.K. Galbraith
|