s, etc., and generally
putting a strain on the filesystem to keep up. It then adds a lot of
overhead to encode that data into 512 byte tar records. The
bottleneck can be hard to see because it's not all I/O, and it's not
all userspace CPU time.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
7;t see where "(too)" would come from.
I know you've forwarded this error before, and I said the same thing.
I can't find anything that would create such a message in
Amanda-2.6.1.
So yes, I think an upgrade is probably the best solution.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
sion, so I'm looking for some
feedback. Those in AUTHORS should feel especially encouraged to speak
up!
The patch is here:
http://github.com/djmitche/amanda/commit/c8c4e7c543bf5bdf93bb87f3a248ffcaab1da82f
As you can see, everything goes except chg-manual, chg-multi,
chg-zd-mtx, and chg-disk.
amgtar -tf - 2>/dev/null | sed -e 's/^\.//'"
which is used to generate the index.
It looks like it happened 1 second after the dump started.
I'm guessing this has to do with your wrapper -- perhaps some kind of
race condition where the wrapper quits before it has read all o
the GNU tar race condition that affects amcheckdump recently,
even though it's been in tar since 1.21. I don't *think* that same
problem would be surfacing here, but it could be.
The index write failed 1s or less after starting the dump. Does the
dump of this filesystem usually ta
as
much as possible, and then we can make some more guesses.
This isn't Nexenta, is it?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
he historical precedent for the content of these log
lines. When we discovered this while rewriting the taper and
reporting code, the immediate temptation was to "fix" it, but then
we'd have a lot of people surprised that the times changed so much in
3.1.1, and I'd be explaining why
mps.
> This is not a ranti btw!
> I'm just surprised that no one who have upgraded to 3.1.x have noticed
> this! Don't people read their amreports :)
If you can boil it down to some numbers that don't add up, then I can
take a look, but I don't have anything concret
, send them to
the list, not just me, so everyone can have a look.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Thu, Aug 19, 2010 at 5:29 PM, Jean-Francois Malouin
wrote:
> The logfile for this run shows:
Can you post the taper debug log, too? We can see what took the extra
48544 seconds, if anything.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Thu, Aug 19, 2010 at 5:50 PM, Jean-Francois Malouin
wrote:
> It's from the time the scribe sets up a transfer (at 22:42:31) to the
> final write (12:11:42 the following day) which is 48551sec, as the
> DONE line in the logfile reports.
>
> You can see me smile in the mirror ;)
That's what I wa
urce is here: http://216.152.245.114/svn/amanda/trunk/ .
I had a look at the patch, by the way - it looks great!
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
d, but I assume you've disabled that.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Thu, Aug 19, 2010 at 6:09 PM, Dustin J. Mitchell wrote:
> There is a bit of taper/driver communication that should go on during
> that pause (REQUEST-NEW-TAPE/NEW-TAPE). That should be recorded in
> the amdump log. Can you attach that?
I got this in private mail (the mailing list
On Fri, Aug 20, 2010 at 1:13 PM, Stefan G. Weichinger wrote:
> Tried that. Big difference.
So try running amdump by hand with the crontab environment.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
s *newline*, not
*whitespace*-terminated.
The patch is simple, so if this explanation makes sense and agrees
with what you see in the patch, please let me know. No need to test
it out.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Fri, Aug 20, 2010 at 8:48 PM, Chris Nighswonger
wrote:
>> http://github.com/djmitche/amanda/commit/z8198
>
> Is it me or is this a broken link?
No, it's me. I totally botched that link, sorry.
http://github.com/djmitche/amanda/commit/z11741
Dustin
--
Open Source Stor
On Sat, Aug 21, 2010 at 6:40 AM, Chris Nighswonger
wrote:
> Removing the call to validate_no_space looks like the correct solution to me.
Thanks! Committed in r3339.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
Like last time, it'd be great if someone could take a look and, if the
patch looks good, let me know. Please feel free to add any comments
on the github commit, too.
Thanks again to Chris for the last review!
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
0001-add-fulls-only-t
atm), it looks fine to me.
Great! Thanks! Buildbot is running amvault, and we'll get some extra
testing in the next releases, so that's good enough for me!
Committed - r3341.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
ore detail, and have a look at the
source to see how it decides whether to generate labels or not. The
logic hasn't changed, but a rebuild may have altered some of the
inputs to that algorithm.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
stitution isn't happening in krb5_init().
Yes, it does. Do you want to fix this up and submit a patch?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
The link should be correct now.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
ot
>
> looks interesting. I miss an example for how the DLEs should look like
> ... and I get errors here .. hmm
I tried to contact the author of this tool, but hey - it's on github.
If someone wants to fork it and fix it up, I'd love to see it merged
into Amanda proper.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
lar to the way Amanda likes to have
read permission to the root of each DLE.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Fri, Aug 27, 2010 at 2:41 PM, Chris Hoogendyk
wrote:
> There was an exchange on the list in the last few months where I asked that
> question and Dustin replied. He said that the historical algorithm had a
> strong preference for an already written tape if one was available that
&
On Tue, Aug 24, 2010 at 2:00 AM, Marc Muehlfeld
wrote:
> Can I use this for migrating hundred of my old DLT 40/80 GB tapes (All
> contain level 0 dumps for long time archiving) to some LTO30 400/800 GB?
Absolutely!
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
27;s analysis of the error on line 1312 is correct.
Can you send along your trace log? I'd like to use it to create a
regression test for this.
DUstin
--
Open Source Storage Engineer
http://www.zmanda.com
On Sun, Aug 29, 2010 at 10:59 AM, Dustin J. Mitchell wrote:
> Fubar means "f*'d up beyond all repair", which I hardly think applies
> to a few warnings! These are minor bugs, which I'll work on a patch
> for. Chris's analysis of the error on line 1312 is cor
the old chg-disk is still not fast-searchable. The new one is.
Old:
tpchanger "chg-disk"
New:
tpchanger "chg-disk:/amanda/vtapes"
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
try running the command I
suggested?
amreport MyConfig --ps=temp.ps
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
's the best way to set the current slot
> now?
amtape CONFIG slot NN
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
amreport immune to this circumstance.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
I'm all for changing error messages - searching the source for error
messages is an underutilized debugging technique!
Dustin
Index: common-src/krb5-security.c
===
--- common-src/krb5-security.c (revision 3343)
+++ common-src
I found the output is
> "ColorLaser." In my amanda.conf I have:
>
> printer "SteveLaser"
>
> Things are just too strange sometimes.
Oh, right, we added a special case in Amanda to print to "ColorLaser"
whenever the "SteveLaser
On Mon, Aug 30, 2010 at 4:13 PM, Tim Nowaczyk wrote:
> Because I'm not a C programmer and I didn't know you could do that. I
> thought my choices were between xstr and strcat/strncat. :)
Gotcha. I'll get the modified version of your patch committed, then.
Dustin
--
On Mon, Aug 30, 2010 at 4:24 PM, Dustin J. Mitchell wrote:
> Gotcha. I'll get the modified version of your patch committed, then.
I meant I'll get it reviewed. Can you review this by trying it out?
http://github.com/djmitche/amanda/commit/krb5-fix.patch
Dustin
--
Open S
e lpr invocation, and the debug logfiles show it passing
along "StevesLaser".
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Mon, Aug 30, 2010 at 2:55 PM, Gene Heskett wrote:
> Dustin; I changed to that syntax and amcheck objected to having both the
> tapedev and that sort of syntax in the tpchanger line unless I was using
> the old chg-disk, so I commented the tapedev line, and gave tpchanger the
> ful
ests for the script (that do not require lvm - mocks are useful
here!).
The 3.2 merge window is closing in ~3-4 weeks, and it would be great
to have this script included in 3.2, even if it has known limitations
at the time.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
system uses the logical AND of the permission bits
from the mountpoint and the root of the mounted filesystem? I don't
know what it does with ownership..
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
it fixed!
I'd poke at the source to see if I could figure that out, but that's
earlier than the Amanda version control goes..
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
pace when deciding whether to
print this warning, but this isn't currently done (that I can see,
anyway).
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
part of the code needs to be tweaked to allow use of any space
> freed by flushed dumps.
Right - that assessment is made at the time, and any dumps that are
not yet flushed are not considered free space. As far as I can tell,
the error is only in printing the warning, not in the actual driv
ctually dumping the same part of the same disk.
To my knoweledge, it is host-specific. Which makes sense, right?
Back in the day when hosts were hosts and spindles were spindles, a
40GB drive on one host was not the same as a 40GB drive on another
host.
Dustin
--
Open Source Storag
l-asym is injecting random
binary data, then that is not correct. Note, however, that trailing
zeroes will cause exactly the same gzip warning. I wish there was a
way to turn this warning off, as the alternative is to frame and
unframe the datastream such that Amanda can identify and remove such
p
o do
with retries, I believe.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
> log.20100903030802.0:25
Can you post the actual grep output? Find is looking for a line of
the format described at
http://wiki.zmanda.com/index.php/Amanda_log_files/Trace_Logs#START_2
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
first line of each of those should avoid the warning you're
seeing. Can you tar up your config, tapelist, and logfiles and send
them along privately?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
I asked him to get into this discussion, but I haven't
seen anything yet.
He could probably use another ping, if you'd like to give it a shot.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Mon, Sep 6, 2010 at 11:59 PM, Dustin J. Mitchell wrote:
> The first line of each of those should avoid the warning you're
> seeing. Can you tar up your config, tapelist, and logfiles and send
> them along privately?
Ah, I didn't realize in your original post that you&
stname or perhaps use a
> wildcarded.domain.name format, Like *.coyote.den there?
You'll have to check the xinetd documentation - I don't know.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
ally
completed, but it needs to be SWIGged and tested and the docs adjusted
appropriately.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
!
Note that I do reserve the right to say, "actually, that's
complicated" (and explain why).
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
d amvault to the
process, there are legitimate reasons to want to run any combination
of {dumper, taper, amvault}.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
in!), but
not quick.
So I'm invoking my "that's complicated" escape clause. The only one
of these that I could conceivably accomplish in the 3.2 timeframe is
#2. I could add 'amadmin $conf hosts' to list all of the hosts,
avoiding the need to grep the multiline
ch or just the tweaked version of
human.pm (from which I can generate a patch)?
I don't mean to put you on the spot - like I said, I can take care of
this if you'd prefer.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
there a problem, or not? What, if anything, should be done about this?
Well, that the changer is failing is a problem. Do you know why
that's happening?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Fri, Sep 10, 2010 at 8:02 AM, Florian Lengyel
wrote:
> Shouldn't amcheck -m DailySet3 send email if there is a changer problem? I'm
> told that amcheck does not send the email.
Yes, and normally it does, so that's either mistakenly reported or
there's a bug of so
7;t say that
> mailto needs those. Perhaps that s/b clarified?
All strings in the configuration file need to be quoted - that's at
the beginning of the amanda.conf manpage.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
Are they unable to use
amrecover? Amfetchdump? Even amrestore?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
st randomly guess that amarchiver would do the
trick, without even looking at the manpage?
It looks like you have some major customer education to do here. Good luck.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Sat, Sep 11, 2010 at 4:22 PM, Jon LaBadie wrote:
> I would like to see a way of saying "do not back more than
> one of these DLEs at the same time".
Duly noted. I don't see a quick way to do that, but I'll create a bug
for it so it's not forgotten.
Dustin
-
ome of those are missing, we could fix that
up.
Do you have a particular example you'd like me to look at?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
catalog from
it (I imagine it would be called amrecatalog). But such a thing does
not exist at this point.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Thu, Sep 9, 2010 at 12:12 PM, Dustin J. Mitchell wrote:
> This is a great idea. I'm actually working on this part of the code
> right now, to make sure that amflush and autoflush will correctly
> flush dumps that are no longer in the disklist, so I think this will
> be relat
On Sun, Sep 12, 2010 at 11:17 PM, Dustin J. Mitchell wrote:
> It includes a rewrite of amdump.sh into amdump.pl. The command-line
> argument handling in shell was just too gnarly to figure out, and
> shell is generally fraught with more portability problems than it's
> worth.
da perl modules, wherever those might
be on your system.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Mon, Sep 13, 2010 at 11:04 AM, Jean-Louis Martineau
wrote:
> Robert, can you try the attached patch?
(only the amoverview hunk is required, right Jean-Louis?)
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
Would that we could wish into existence an army of volunteer
developers!
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
component you quoted
is correct - it's the options for that host.
What are you seeing that's got you thinking the taper never gets any
data? Can you take a look at the taper debug log and see what you
find?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
ng to
become gnutar-specific and start futzing with the data in the backups
on the server side.
As you can see, complicated. But a consistent approach to storing the
DLE and path of a particular "user object" over time would be a useful
first step. Do you have any thoughts on how that might be
implemented?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
(starting a new thread)
On Mon, Sep 13, 2010 at 9:16 AM, Gunnarsson, Gunnar
wrote:
> Hi Dustin,
>
> A script-tool application for running script taking parameters such as path
> to script, PFEXEC or SUDO and when to run etc
>
> define script-tool amplugin-script {
> plu
d add some
additional code to amdump to send the next-tape email.
The chg-human thing needs to be done eventually, but demand has been
low and there are lots of tough design questions to answer first.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Mon, Sep 13, 2010 at 1:44 PM, McGraw, Robert P wrote:
> I thought it was fixed but this time the warnings were at the beginning of
> the output and I did not catch them.
Jean-Louis' patch only addresses one of the warnings. Try applying
the patch I linked to, instead.
Dusti
ed,
but as it stands there are some *very* significant unanswered
questions, and probably a lot of more subtle problems, too.
Instead, we should look at how other backup software handles similar
problems, and consider throwing out some long-standing Amanda
weakne^Wfeatures. You can see why this becomes
ance you or Stefan could do a fresh install from the
updated ebuilds (I believe robbat has just put 3.1.2 in ~x86) and see
what problems you come to? They'll make good gentoo bugs for the
ebuild, but also a good concrete symptom that we can look at fixing in
Amanda.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
ed that the long options with a single dash work at all,
but the "only need a unique prefix" thing is well-known. I'll get the
docs and usage fixed up to use double-dashes for long options.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
u
tell what it was doing via gdb or strace?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
mand in gdb to figure out
what threads are active, and then the 'thread' command to switch to
each one and the 'backtrace' command to generate a backtrace for it.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
t.
How can we fix that? Jean-Louis has most of the knowledge required,
but we're both busily working on the upcoming 3.2 release, so we'll
need someone else to take the lead in getting that documentation put
together. I'm not sure if it should be in a manpage or on the wiki.
Dusti
These are some exciting ideas, so far!
I'd love to see some of this fleshed out in code, even if that's not
Amanda code -- an example of a catalog with fake dump and recover
operations that demonstrates how all of this would fit together. With
that in place, we can settle on a particular structur
uot;feature" of Perl's
GetOpt::Long. Let's just pretend it doesn't exist :)
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
ch will probably still have the warnings that Jon noted.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
any case. I'm not at all familiar with msmtp..
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
g vociferously that Amanda's throughput "sucked", when in
fact their network had a high packet loss rate because their server
was hooked into a hub with a massive number of collisions. Hopefully
your situation isn't quite that dire..
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
ormatting, but it looks like you're misusing awk and just printing
amadmin commands. You need to actually *run* them.
The backupcentral forums are very badly translated into the
amanda-users list and back. It would be easier for us to help you out
if you posted directly to the list:
http://ama
mped the whole command
into a single string and thus let the shell break them at space
boundaries.
I'll have a patch for your experimentation in a little bit.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
be a
> 64-bit problem. Any insight would be highly appreciated!
I don't know much about the various tar formats, but I do know that
GNU tar supports a few variants, some of which work better with very
long filenames than others. Is your wrapper forcing one particular
format? Can you
On Thu, Sep 16, 2010 at 12:40 PM, Stefan G. Weichinger wrote:
> nice to hear that!
Give this a try:
http://github.com/djmitche/amanda/commit/z11963.patch
Don't worry about the installcheck hunk. You should be able to apply
the amreport hunk in place in /usr/sbin if that's eas
taper has
provided. Can you check the taper debug logs to see if it's failing
for some reason, and also check the amdump logs to see what the
taper's reply to the PORT-WRITE command looks like?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
hing to report on!
>
> if I want to force a mail sent ...
--from-amdump is really only meant to be used, well, from amdump :)
If you just run 'amreport daily' it will send a report to stdout
If you want to mail the report, use
amreport daily --mail-text="f...@bar b...@baz"
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Thu, Sep 16, 2010 at 2:14 PM, Stefan G. Weichinger wrote:
> Tested, works now!
Great! Jean-Louis just reviewed the patch, and it's committed in r3411.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
elease will
be.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Thu, Sep 16, 2010 at 4:14 PM, Dustin J. Mitchell wrote:
> That would be great. We're rapidly heading toward a 3.2.0b1, and the
> more testing can happen with that, the better the 3.2.0 release will
> be.
In fact (and sorry for the double-post .. I've probably had too much
hosted at overlay-servers, I have to research
> that ...
Lisa, do you have any pointers on that front?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
On Thu, Sep 9, 2010 at 11:05 PM, Dustin J. Mitchell wrote:
> I don't mean to put you on the spot - like I said, I can take care of
> this if you'd prefer.
Are you working on this? Perhaps someone else on the list can jump in?
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
changes.
As for the automake errors - if you look at the autogen script
http://github.com/zmanda/amanda/blob/master/autogen
You'll see it does a number of things before calling aclocal and the
other autotools. You'll need to replicate that in the svn ebuild.
Dustin
--
Open Source
ut mode, although it's not tested at all so
currently YMMV. But that might be a nice way to start to export
status data from Amanda into a tool designed for monitoring and
notification.
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
#x27;m no gentoo dev, but it builds and installs and passes amcheck for me.
Now, if it can get put in an overlay, even cooler!
Dustin
--
Open Source Storage Engineer
http://www.zmanda.com
1101 - 1200 of 1235 matches
Mail list logo
