es appear to be produced.
>
> I would really appreciate some help on determining what is happening as I can
> not fix it until I do.
Generally this kind of intermittent failure is caused by timeouts that
are too low. Try increasing the timeouts.
http://wiki.zmanda.com/index.php/Result
n, and investigate the contents of the 'mt' struct. As
is always a frustration with hardware-related questions, this is not
something anyone can investigate on your behalf :(
Another solution may be to just delete lines 537-544, recompile, and
see what happens.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
n of
> why it didn't run or any error it might have
Is there a correlation to dump level?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ouch all of the finished manpages
so that the Makefile isn't tempted to rebuild them.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
the time!
> Running with 30 virtual tapes on a 372GB sata drive now. Bare metal recovery
> could be problematic, and would require a bare bones install that supported
> sata since my bios does not. But that obviously is not an amanda
> problem. :-)
We'll survive without t
ning.
Platform Experts, this means you! Stand up and be
counted!
Also, I wrote up a thorough guide to running the full suite of tests here:
http://wiki.zmanda.com/index.php/Testing
Please let us know what you find -- positive or negative.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
C8 font problem, or at least a
gnuplot problem. If you run one of the gnuplot examples
(http://www.duke.edu/~hpgavin/gnuplot.html), I expect you'll have the
same problem. If not, then maybe there's a problem in amplot.gp.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
r the changer is mixed up.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
exec?
It is in the NEWS and ReleaseNotes files, but not yet on the wiki
(since it hasn't been released yet).
Dustin
--
Storage Software Engineer
http://www.zmanda.com
This is getting to be a FAQ. See
http://marc.info/?l=amanda-users&m=120298649014627&w=2
At the moment, there's no easy way to add arguments to gtar.
Downgrading tar seems to work, though.
Dustin
P.S. A dozen cookies on me for the person to write this up on the wiki :)
On Fri,
houldn't have a problem with either of those situations.
Amanda stores the source application and information about any filters
in the dump header, so as long as the application and filters still
exist, you'll be able to recover your data.
But don't take my word for it --
> What would be the correct syntax to use in determining how many lines of code
> there is in the current 2.6.0b2 tarball?
If you don't have sloccount:
Total Physical Source Lines of Code (SLOC)= 121,066
Development Effort Estimate, Person-Years (Person-Months) = 30.78 (369.30)
(
under active development. For a hint as to
how complex this is, see
http://wiki.zmanda.com/index.php/User:Dustin/Data_Handling_Model.
There's no specific ETA on this development at the moment.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
> Could not find any valid dump image, check directory.
>
> but all the host._disk.part is in the holding directory.
What is the output of
find /Data/amanda/amanda-hold -type f -ls
Dustin
--
Storage Software Engineer
http://www.zmanda.com
may be that something funny happened
with the filesystem -- perhaps the directory gtar was working in was
deleted? Was NFS involved? ENOMEM tends to be used for reasons other
than its literal meaning.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ng/file bs=1024 count=1
to look at the header for each holding file, and see if anything looks amiss.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
e bio
> server for something that was backed up on the geo server. All I need is
> the tape.
I did this for a long time between two colo'd servers at ThePlanet.
FWIW, I also rsync'd the holding disks, so data that was still
"precariously" on the box from which it ha
mit last night
that wasn't fixed before the snapshot was taken. The commit is fixed
now.
Still, good to know you're out there, watching us like a hawk.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
-1_left1_S00025L31 158 OK
> > 2008-03-03 19:10:01 bigbrain BIG_BRAIN_pm3903_postmortem_tif 0
> av24-1_left1_S00025L32 159 OK
> > 2008-03-03 19:10:01 bigbrain BIG_BRAIN_pm3903_postmortem_tif 0
> av24-1_left1_S00025L33 160 OK
And again -- a level 0 (in three parts) on tape 25.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
any further difficulties. Also, be sure to keep your eye on
upcoming releases, and install the 2.6.0 release when it is ready.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
>counter == 0' failed.
> sendbackup: time 841.420: index created successfully
> sendbackup: time 841.423: error [dump (31206) /bin/tar got signal 6]
> sendbackup: time 841.423: pid 31202 finish time Wed Mar 12 05:23:05 2008
Definitely a bug in tar -- it called abort() when the assertion
patched version of that utility
and send it to you by email.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
On Thu, Mar 13, 2008 at 8:32 AM, Jean-Francois Malouin
<[EMAIL PROTECTED]> wrote:
> Thank you Dustin.
>
> Anyone is aware of issues with gnutar 1.16 and amanda?
This page:
http://wiki.zmanda.com/index.php/FAQ:What_versions_of_GNU_Tar_are_Amanda-compatible%3F
lists a
ug-tar -- subscribe
http://lists.gnu.org/archive/html/bug-tar/2008-03/msg00023.html --
Jean-Louis' patch post
http://www.mail-archive.com/[EMAIL PROTECTED]/msg01767.html -- Jordan
Desroche's post
google can help you find some more posts.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ould inspect and repair the index
> files when this happened? Probably lost when I woke up one morning and found
> my well developed FC6 install wasn't re-bootable, LSN0 on /dev/hda had one
> non-zero byte left in it.
Yep -- it's called tar-snapshot-edit, and it
post the two
commmand lines, one that works, and one that doesn't. We'll try to
whittle them down to see what the difference is.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ot --one-file-system \
--listed-incremental /root/gtar_test/gnutar-lists/172.31.3.9_boot_0.new \
--sparse --ignore-failed-read --totals . \
| wc -c
gtar --create --file - --directory /boot --one-file-system \
--sparse --ignore-failed-read --totals . \
| wc -c
gtar --create --file - --directory /boot --one-file-system . \
| wc -c
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ontrol-c
- zip up /tmp/trace and sent it to me (not to the list -- I'll summarize)
Dustin
--
Storage Software Engineer
http://www.zmanda.com
On Fri, Mar 28, 2008 at 5:57 PM, Oscar Ricardo Silva
<[EMAIL PROTECTED]> wrote:
> /usr/local/amanda25/sbin/amrecover -C dailytoo
> AMRECOVER Version 2.5.2p1. Contacting server on amanda.tn.utexas.edu ...
> [request failed: amanda.tn.utexas.edu: ssh could not resolve hostname]
...
> It's not re
rouble :(
Anyway, thanks to your speaking up there's a patch in process to make
the original message more informative:
WARNING: snarsblat: selfcheck request failed: ssh_security could not
find canonical name for 'snarsblat': Name or service not known
Dustin
--
Storage Software Engineer
http://www.zmanda.com
16) = 9216
write(1, "./\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
10240) = -1 EPIPE (Broken pipe)
--- SIGPIPE (Broken pipe) @ 0 (0) ---
+++ killed by SIGPIPE +++
I'm running tar 1.19, and also tested this against 1.15 and 1.13.25,
and saw similar (that is, correct) behavior everywhere. What version
of tar are you running?
Dustin
exhibit this problem, and 1.13.25 has been
tested to work properly with Amanda. All of the virtual machines I've
ever used were manually installed, so I have no idea what a VMware
distro would include.
Could you try
strace -f tar -cf - . 2>/tmp/trace | true; tail /tmp/trace
and
n't need to be the same everywhere -- you can add 'auth
"bsdtcp"' to the DLEs for this client:
hostname /path/to/dle {
user-tar
auth "bsdtcp"
}
This would be an excellent diagnostic check, at least. If it works,
then you know you have a firewall probl
tar ouput?
This page should be able to help:
http://wiki.zmanda.com/index.php/TCP/UDP_ports
Dustin
--
Storage Software Engineer
http://www.zmanda.com
m has glib2 at all, it
likely has a new enough version.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
rce of glib and compile it 64bit.
> Any hints?
It's an easy, knob-free compile.
We've been arried by bit-length mismatches too. Sometimes I wish
multilib had never been invented, and we all had been told, "32- or
64-bit. Those are your options."
Dustin
--
Storage Software Engineer
http://www.zmanda.com
iously like amrecover isn't even trying to decrypt
the data. The funny thing about Amanda is that all of the
"unfiltering" of data on restore is done on the server, regardless of
the client/server settings in the DLE, and happens through
'amfetchdump'. So a good pla
x27;s a vestige of a previous
> encryption framework?
The quick answer is that amrecover has to "pass through" encryption
information from amindexd and back to amidxtaped. So even though it's
not doing the decryption, it needs to know about it.
I can change the message, if you&
Thanks -- I fixed it on the wiki, and I'll patch the codebase in a bit.
Dustin
On Fri, Apr 4, 2008 at 3:37 PM, Oscar Ricardo Silva
<[EMAIL PROTECTED]> wrote:
> Since I've been this deep (holds hand above head) in trying to get
> encryption to work, I've been reading a
ailable" corresponds to EAGAIN, and is
usually seen with non-blocking IO, which Amanda doesn't use, so I'm
not sure what to make of that information. The dumper debug log may
clarify that.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
Any idea what 127.0.0.2 is doing in there?
Do you mind emailing me a selection of full debug files for the last
two or three runs, with an indication of which dumps failed? I'll see
what I can find.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
file shell function.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
m is, the answer is probably going to be "upgrade to 2.6.0".
Dustin
--
Storage Software Engineer
http://www.zmanda.com
o rebuild Amanda with a bigger range of ports
or (if you're feeling hackish) to remove the other services in that
range from your /etc/services. I'll look into changing the error
message.
HTH
Dustin
--
Storage Software Engineer
http://www.zmanda.com
On Mon, Apr 14, 2008 at 4:51 PM, Charles Stroom
<[EMAIL PROTECTED]> wrote:
> Dustin, thanks for the analysis and is seems to make sense because
> indeed in the /etc/services only 703 and 708 are unassigned in between
> 700-710. However, my amanda is installed from rpm from the d
" (", qdisk, "): ", strerror(errno), NULL);
827 }
where 'amode' is R_OK. You can simulate this with:
perl -MPOSIX -e 'print "yes\n" if POSIX::access("/", &POSIX::R_OK)'
Dustin
--
Storage Software Engineer
http://www.zmanda.com
Hmm. Can you use 'strace' to see what it's up to? Is the changer
doing anything interesting?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
e changer through a number of slots.
If you use the '-f' flag to strace, it will trace the children, too.
You should be able to see an 'exec' after the clones. It will be a
lot of data, but it's not too hard to search through.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ne this, please post your wrappers and/or add them to the wiki?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
If you're not committed to 2.5.0p2, then upgrading to 2.6.0 (or at
least 2.5.2p1) would be better.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
. 2.6.0 is basically the same to 2.5.2p1 as far as
compatibility is concerned.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ot sure what you mean by 'dump' -- vtapes? holding? logs?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
the value of the config
parameter "usetimestamps"?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
you provide new/changed text, I'll be glad to add it!
Dustin
--
Storage Software Engineer
http://www.zmanda.com
-f daily && mt offline' | at now
Thanks! It's in review, and should be committed shortly.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
heard boo about it. I'll see if I can replicate it. I'm
happy to have more eyes on the code!
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ee whether the REP packet or its ACK is actually being
lost.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
Amanda 2.5.0 on your server. Otherwise it just gives Amanda better
time resolution -- and fixes this bug. I hope :)
Dustin
On Mon, Apr 28, 2008 at 7:53 AM, julien brule <[EMAIL PROTECTED]> wrote:
> hi all,
>
> Does someone have an idea for flushing my chunk to the tape when you g
nda.com/index.php/Splitting_dumps_across_tapes
for the skinny on setting it up.
> I'm testing a newer version, and was just wondering...haven't really touched
> Amanda for more than a year because it never breaks...
Amanda's great that way!
Dustin
--
Storage Software Engineer
http://www.zmanda.com
e spanning enabled?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
On Tue, May 13, 2008 at 1:54 PM, Chris Marble <[EMAIL PROTECTED]> wrote:
> Last I looked Amanda did not make use of the speed parameter (but that
> was many years ago). Not sure why my speed number is so much higher.
This is still the case.
Dustin
--
Storage Software E
large enough to accomodate the parallelism used in Amanda,
the only implication of these ranges is for firewall rules.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ve, and use
chg-multi. As long as the drives are swapped and mounted such that
tape N is online when Amanda wants it, all should work well.
You may also want to do an 'amadmin force' at the beginnning of each
week to make sure that you get level 0's on each drive.
Dustin
--
ports used for reverse connections back to the server. As long as the
>> ranges are large enough to accomodate the parallelism used in Amanda,
>> the only implication of these ranges is for firewall rules.
>>
>> Dustin
>
> So, is that a "no, the clients need n
ailing list where I could ask this question
- an HP/UX box I can play around on for a little while
Thanks!
Dustin
--
Storage Software Engineer
http://www.zmanda.com
36203.c
Description: Binary data
#! /usr/bin/perl
use warnings;
use strict;
use Socket;
use Sys::Hostname;
my $iaddr = gethostby
ing *-hp-hpux*?
The attached patch does just this. Scott, can you give it a whirl?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
hpux.patch
Description: Binary data
/Users/amanda/.amandahosts: Permission denied"
>
> Anyone have any ideas?
Doug Kingston reported a different incarnation of what I believe to be
the same bug[1] -- please try amanda-2.6.0p1 and let us know if you
still see the problem. You can probably also fix the problem by
removi
use that file listing as an index, as thats exactly how Amanda
builds the index.
There's currently no neat amanda command to do this, although it is
planned as part of the Application API (the "reindex" operation).
Dustin
--
Storage Software Engineer
http://www.zmanda.com
s a different bug than
that reported by Doug.
If you still have the source directory handy, do you mind applying the
attached patch, and then sending the amandad.* logfile generated (on
the client) during a run of 'amcheck'? You can find the right logfile
by grepping for "pw_uid-A". Please send the logfile along even if the
run succeeds.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
yeatman.patch
Description: Binary data
ere a way to examine outstanding filesystem locks on AIX, to see
if there is a process holding a lock on this file?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
nt out the code
that is locking amandates, on the assumption that you are not using
it.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
or won't unlock)
> amandates ?
Possible .. there's been quite a bit of work on the locking code since
the 2.4 series..
> Is there a howto for AMANDA and AIX. Not many folks still using AIX
> these days ...
http://wiki.zmanda.com/index.php/Installation/OS_Specific_Notes/Installing_Amand
FLAGS=-L/opt/local/include
>
> Hmmm, what's going on here?!
You'll need to use --with-amlibexecdir=/usr/local/libexec. Sorry about that :)
Dustin
--
Storage Software Engineer
http://www.zmanda.com
bsdudp' driver" is specified in the plist file,
> right? Why am I lacking a "driver of security type 'bsdudp'"?
You probably didn't build it. I would strongly recommend that you use
bsdtcp, anyway.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
o add
auth "bsdtcp"
to the dumptypes you're using (probably easiest to just add it to the
global dumptype).
Dustin
--
Storage Software Engineer
http://www.zmanda.com
lient in /etc/amanda/amanda-client.conf.
The client should be listening for TCP connections using the bsdtcp plist.
It's messy, I know -- it's all there for historical reasons.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ions
-- I've not been paying attention to this thread.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
vious tape out of the drive, stick in a new one,
and Amanda picks right up where it left off.
VOLUNTEER OPPORTUNITY:
The changer documentation on the wiki could use some help. If you
use chg-manual, head over to http://wiki.zmanda.com/index.php/Changers
and make it better!
Dustin
--
Storage Software Engineer
http://www.zmanda.com
fixed. Amanda has a bad
habit of storing some lengths in kb and some in bytes. We're slowly
standardizing on bytes.
As for the rest -- I don't have an immediate answer. Can you send
your whole amflush log, as well as the taper debug log?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
king, it hasn't
been fixed in the 2.6.0 branch. I'll take care of that before the
next patch release.
Your fix -- touching the file -- should be adequate until then.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
On Sun, Jun 8, 2008 at 11:13 PM, Dustin J. Mitchell <[EMAIL PROTECTED]> wrote:
> This is a result of an error-messaging bug in the 2.6.0. Basically,
> config error messages go to stderr, which xinetd routes to the server.
> The bug has been fixed in trunk, but on double-chec
since 2.5.1 is no longer maintained, the solution is to stop tickling
the bug :). See if you can figure out why that connection is timing
out -- busy network? Downed client? Network partition?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ick on the link "nominate this project in another
category."
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ou could try hacking chg-manual a little bit to send its messages to
e.g., /var/log/chg-manual.log, and then just "tail" that to see what
it's up to.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
ere's a rather flexible C module for analyzing the holding
disk. If it would help, I'd be happy to talk about SWIGging that into
perl, allowing your script to interface with it directly.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
to holding-disk only
amdump -otapdev= -otpchanger=
# backup to tape
amdump
Then yes, that is correct.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
remain the same. The interface might change somewhat in that case.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
On Thu, Jun 19, 2008 at 2:42 PM, Dustin J. Mitchell <[EMAIL PROTECTED]> wrote:
> Take a look at server-src/holding.h. Even if the C is not your
> flavor, the comments give a pretty good description of how things get
> broken down.
Sorry, here's a link:
http://amanda.svn.so
catalog. Amanda should be cleaning
out the logs that correspond to overwritten tapes. Do you have tapes
that old?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
On Fri, Jun 20, 2008 at 3:27 PM, Paul Yeatman
<[EMAIL PROTECTED]> wrote:
>driver: result time 35026.575 from dumper0: NO-ROOM 00-00016 0
>
> Not sure what went so awry with this run. It is likely back from when
> Dustin was working with me on upgrading to 2.6.0 and
olks
like you with lots of tapes won't have to worry about manually
manipulating this stuff. And hopefully we can be a bit more succinct
in how the catalog is stored.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
hen manually
> untar all the restored chunks?
If a recover requires multiple tapes, then amrecover will ask for
them, either via the changer script (I've forgotten whether you're
using chg-manual or not) or manually.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
recommend the first option.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
27;t seen this happen more often. This is
the first time in my memory, but I haven't been around that long :)
Do you want to send along a patch?
Dustin
--
Storage Software Engineer
http://www.zmanda.com
on
> my system.
We had a bit of this sort of trouble in 2.5.2, and the code has been
re-worked in 2.6.0. I imagine that some rather rough edits to the
source (removing #include from amanda.h) will fix the
problem.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
POSIX support, and Xenix is basically gone at
this point. I assume UnixWare supports the POSIX standard.
Does anyone out there intend to install the Amanda server, version
2.6.0 or higher, on hardware that does not support the POSIX tape
interface? If not, I'd like to trim the fat..
Dustin
-
On Fri, Jun 27, 2008 at 4:22 AM, Weiwei Luo <[EMAIL PROTECTED]> wrote:
> May I know which setting make amcheck to check the directory at /home/amanda?
It's the home directory of the 'amandabackup' user on the client (or
whatever CLIENT_LOGIN is set to).
Dustin
--
Storag
due to point 1.
Nice job investigating this, by the way! If you want to stick around
and hack on Amanda, we'd love to have you.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
, and don't have
a test environment that uses kerberos, so it will be good to have
someone around to help out.
Dustin
--
Storage Software Engineer
http://www.zmanda.com
;d prefer that you get this building by
whatever means necessary and move on to packaging 2.6.0! :)
Dustin
--
Storage Software Engineer
http://www.zmanda.com
201 - 300 of 1163 matches
Mail list logo