> "CH" == Chris Hassell writes:
CH> Most build systems these days recommend building entirely as a
CH> normal user (even Debian?) and the chown-and-other-ops are done with
CH> careful setuid privileges given to the build system.
CH> RPM and rpmbuild have been doing it that way for years.
> "BC" == Cuttler, Brian R (HEALTH) writes:
BC> Can anyone provide the files I need, give hints on what I'm doing
BC> wrong or help in some other way?
I don't believe Amanda upstream ships systemd unit files, so you would
need to provide them yourself. Perhaps look at the files Fedora
I'm trying to understand what's supposed to happen when amrecover gets
to the end of the backup. Currently I'm running amanda 3.5.1 as
packaged by Fedora. (We only lightly patch it, and as far as I can tell
none of the patches should be relevant to amrecover.)
Basically, I do a simple recovery
> "GK" == George Kelbley writes:
GK> My Department also purchased the Enterprise Zmanda license to get
GK> NDMP support. However our lastest upgrade on the file server changed
GK> the NDMP requirements and Zmanda can no longer use it. I've
GK> switched to have Zmanda use
> "JM" == Jean-Francois Malouin
> writes:
JM> This is not was I have experienced, at least with 3.3.x.
For the record, my server is 3.5.1.
JM> I've been using the following in some configs with the expected
JM> result:
OK, it's good to know
And so it turns out that if you turn off autoflush, Amanda will never
flush existing dumps to tape regardless of the flush-threshold-*
settings. Which I guess makes sense. And with "autoflush yes", it
seems to simply flush everything currently on the holding disk
regardless of other settings.
> "DSB" == Debra S Baddorf writes:
DSB> Well, if it’s already not doing what you want, you might read
DSB> these.
Well, yeah, that's pretty much the same as I indicated that I'm using
now, except that I'm using different thresholds. My problem is that I
asked for at least
With tape sizes what they are now I'm finally at the point where I have
way more tape than stuff to back up. To satisfy my inherent laziness,
I'd like to actually fill every tape (or at least come close to filling
them). And then just leave my library full of tapes without ever
needing to change
> "HJ" == Henrik Johansson writes:
HJ> Are there any known release-plans for amanda on CentOS-7?
That is up to Red Hat (since CentOS almost certainly won't diverge from
them in this instance). They will not disclose any future plans, and
honestly I doubt they
> "HJ" == Henrik Johansson writes:
HJ> How old are version 3.3.3?
It's pretty old, but Red Hat is incredibly conservative regarding
updates. It was released on January 10, 2013 but Red Hat has almost
certainly applied multiple patches to it.
HJ> The version
Thanks for posting that. I've very rarely seen errors like those myself
but haven't invested the time to track them down.
Unless 3.4.6 is imminent, I will go ahead and add this patch to the
Fedora packages.
- J<
> "JMC" == Jose M Calhariz writes:
JMC> I happily ship a systemd unit for amanda if someone writes one for
JMC> the Debian package.
Amanda in Fedora has been socket-activated for years. Feel free to grab
the relevant units from our git tree:
Sorry to keep replying to myself, but after 300MB of debuginfo, here's a
more complete backtrace:
Core was generated by `/usr/bin/perl /usr/lib64/amanda/amcheck-device math
math'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 strlen () at ../sysdeps/x86_64/strlen.S:106
106
To verify everything, I swapped back in a regular tape, no problem. I
put the cleaning tape back in and the problem returns.
Checking coredumpctl (this is a Fedora system, so systemd is handling
the cores) I see that... perl itself is segfaulting. Looking at the
backtrace (which is
I replaced the last tape in my 48-slot library with a cleaning tape
since I'd rather not bother with manually cleaning the drive. But this
appears to make amcheck segfault:
AMANDA:backup2:amanda/server/math> amcheck -s math
Amanda Tape Server Host Check
-
NOTE:
In https://bugzilla.redhat.com/show_bug.cgi?id=1423471 a Fedora user
reports that amdump simply does nothing in 3.4.2 besides return exit
code 9 (error || RESULT_MISSING). However, 3.4.1 works fine. We didn't
change anything in the packaging between 3.4.1 and 3.4.2 besides add the
patch to fix
> "JM" == Jean-Louis Martineau writes:
JM> Henrik, amanda 3.4 create shared memory and semaphore and require RW
JM> on /dev/shm. You should report it to fedora since they amanda
JM> policy.
Fedora already knows, but sadly I just don't know what's going on with
the
> "JM" == Jean-Louis Martineau writes:
JM> Try the attached patch.
That does appear to work fine for me. I'm getting good backups of the
my old hosts now.
Thanks for your help. I will push these two one-liners into the Fedora packages
until there's new point
> "JM" == Jean-Louis Martineau writes:
JM> Removing the leading semicolon was a mistake. Your patch looks
JM> good. let me know if it works so I can commit it.
Sadly my backup run didn't succeed, but I doubt that match is involved.
Instead there appear to be
> "JM" == Jean-Louis Martineau writes:
JM> Removing the leading semicolon was a mistake. Your patch looks
JM> good. let me know if it works so I can commit it.
Of course; we'll see tonight.
JM> I have no setup to test older version and they don't compile on
JM>
> "JLT" == Jason L Tibbitts writes:
JLT> The problem seems to be due to a change in the OPTIONS which the
JLT> server sends, from:
The quickest hack I found was this:
diff --git a/server-src/diskfile.c b/server-src/diskfile.c
index ffee14e..929c05e 100644
---
I just updated my amanda server from 3.3.9 to 3.4.1 and have found that
I can no longer backup a RHEL5 client running the stock RHEL5 amanda
2.5.0p2 package.
The problem seems to be due to a change in the OPTIONS which the
server sends, from:
DUMP /local/samba 0 OPTIONS
> "jc" == jon@jgcomp com writes:
jc> Will try. I was still researching and hoping for some alternative,
jc> actual fix. Seems not to be an unusual situation.
That's going to have to come either from the selinux policy authors or
from someone who sits down and learns
> "jc" == jon@jgcomp com writes:
jc> Ok, I confirmed my home dir can be backed up with selinux set to
jc> non-enforcing.
How about just setting amanda_t to permissive as I suggested in my
previous message? At least then you wouldn't have to disable selinux
throughout your
> "UM" == Uwe Menges writes:
UM> I filed https://bugzilla.redhat.com/show_bug.cgi?id=1280526 but I
UM> don't know if this is really the same cause.
Sadly that one got closed because the release was never updated to
something newer than Fedora 22. I suspect it's still a
> "JLT" == Jason L Tibbitts writes:
JLT> And that's what I did and it shows the problem pretty clearly:
JLT> xfsrestore is segfaulting. In fact, xfsrestore segfaults trying to
JLT> restore any of the backups of this filesystem.
I should really be fair to the XFS devs
> "JLT" == Jason L Tibbitts writes:
JLT> Are there any tricks for trying to duplicate the index generation?
JLT> I guess I could try to pull a raw dump file off of tape and run the
JLT> same xfsrestore command line and see what happens.
And that's what I did and it shows
uggestions anyone might have.
--
Jason L Tibbitts III - ti...@math.uh.edu - 713/743-3486 - 660PGH
System Manager: University of Houston Department of Mathematics
Tue Dec 13 00:12:12.277433734 2016: pid 27392: thd-0x55832f61cc00: sendbackup:
pid 27392 ruid 33 euid 33 version 3.4.1: start at Tue De
> "SGW" == Stefan G Weichinger writes:
SGW> I assume the missing labels will be scanned step by step when the
SGW> tapes are used and rotated? Or do we miss some step?
It's been long enough since I've added new tapes that I've forgotten how
it works. I think I simply
> "SGW" == Stefan G Weichinger writes:
SGW> If the admin forgets to change tapes every week and the next tape
SGW> to use is NOT in the library, amanda goes wild and doesn't stop
SGW> scanning for the next tape
You don't actually say what kind of library you have so you
> "TR" == Tom Robinson writes:
TR> Agreed. The dump utility has a long history. Looking at the
TR> changelog on the CentOS 7 rpm package and the sourceforge bug list I
TR> can see that Red Hat have been active in maintenance.
You can see the patches which Fedora
DSB == Debra S Baddorf badd...@fnal.gov writes:
DSB O. OK, next answer might have to be from an expert. I’m
DSB guessing that amanda thinks “well, maybe he already DID put another
DSB tape in a slot somewhere, and I just have to go look for it”. I
DSB *would* think it would scan the
JM == Jean-Louis Martineau martin...@zmanda.com writes:
JM Which taperscan are you using?
Heh, I had no idea those even existed. I've been porting my amanda.conf
around since, well, a very long time ago, and it didn't have a taperscan
definition.
JM The default 'traditional' read all slots.
DSB == Debra S Baddorf badd...@fnal.gov writes:
DSB Are you using the “robot” type changer?
Yes:
define changer magnum {
tpchanger chg-robot:/dev/sg8
property tape-device 0=tape:/dev/nst0
device-property LEOM TRUE
device-property BLOCK_SIZE 512k
changerfile
I'm running 3.3.6 (but could easily upgrade if it would help). For a
long time I've been wondering about amanda's behaves when I have a
tapecycle larger than the number of tapes in my barcode-supporting
library and all of the tapes in the library have been recently written.
When the next tape in
JM == Jean-Louis Martineau martin...@zmanda.com writes:
JM The main problem is that if you leave the dump in the holding disk,
JM amanda will automatically re-flush (autoflush) them on the next run.
JM There is no way to store the information about dump that are already
JM flushed and dump that
My amanda server has a really large holding disk, because disk is cheap
and because lots of disk striping generally equals better write
performance.
The usual restore operation I have to do is pulling things off of last
night's backup, which involves waiting for the library to load things
and
MS == Michael Stauffer mgsta...@gmail.com writes:
MS Hi all, I'm setting up a new amanda server. Am I right in seeing
MS that RHEL/Centos 7 is not yet supported? If so I'll just install
MS Centos 6 as before. Thanks.
Why do you think it wouldn't be supported? Just yum install
amanda-server.
Just checked; my amanda run that started at 12:05 Sunday morning
finished at 3:15AM. No problems at all. I'm running 3.3.6 as packaged
by Fedora. Pretty sure it has no patches relating to this.
- J
This seems like it should be easy, but I've not been able to figure it
out. The current RHEL7 version of Amanda is (intentionally) built
without dump installed. This leads to amcheck failing with:
ERROR: nas00: [DUMP program not available]
ERROR: nas00: [RESTORE program not available]
Is it
PH == Petr Hracek phra...@redhat.com writes:
PH Can be amanda uninstalled (2.6.p1) and again installed (3.3.3)?
PH What about configuration file or database changes?
I don't think that's possible. I installed a new machine and just
copied the configuration files over (so no idea about the
Has anyone experimented with using ionice to bump up the priority of the
taper processes? I have a pretty solid amanda server, but the LTO4
drive I've recently started using is so fast that I don't think I'm
keeping it saturated. I'd happily trade off dumper performance to keep
the tape drive
I have recently acquired a 24-tape LTO4 library which has been running
fine with amanda 2.6.1p2 as packaged as part of Fedora 13. I've just
had to change out the initial load of tapes tapes for fresh ones. I've
labeled them all without incident, B00025 through B00048. amadmin
tape indicates
CH == Chris Hoogendyk hoogen...@bio.umass.edu writes:
CH Temporarily change your tapecycle to the total number of tapes you
CH have (including all the new ones).
I have 48 total tapes labeled (24 currently in the library), so I set
tapecycle to 48. Regardless of which tape I load, amcheck will
DL == Dominic Lepiane dominic.lepi...@ptgrey.com writes:
DL Dear list, We have recently acquired a Dell PowerVault TL2000 tape
DL library. It has an LTO-4 drive, 2 magazines with a total of 24
DL slots including an import / export slot on slot 24:
You can't generally pick from the I/E slot to
4 == 440 dr...@ihavefire.com writes:
4 Is there a way to get a backup to remain in the holding disk?
I agree this would be nice. I'd go even further and wish that I could
keep several nights' backups on the holding disk, even to the point of
keeping it somewhat full and removing old backups
RW == Roger Williams r.willi...@gns.cri.nz writes:
RW I have been very disappointed to discover that the performance of
RW Linux dump seems to be completely awful in recent RedHat Enterprise
RW and CentOS releases.
RW Have other Amanda admins seen this?
Yes, I've seen the same. I've no idea
?
--
Jason L Tibbitts III - [EMAIL PROTECTED] - 713/743-3486 - 660PGH - 94 PC800
System Manager: University of Houston Department of Mathematics
JL == Jon LaBadie [EMAIL PROTECTED] writes:
JL I'm not sure what you were expecting amcheck to do.
I was expecting it to alert me to the fact that no suitable tape was
found and that I should stick a new tape in the changer. If I don't
change out tapes, the backups tonight will fail. If I run
lw == lone wolf [EMAIL PROTECTED] writes:
lw 2) Has anyone used the SDLT 320 drives w/ Amanda? I searched the
lw group archives and didn't get much from them on these drives, same
lw with the FAQ (tape-info would be nice for these drives).
I have an SDLT220 drive; it works just fine. The only
BH == Brook Hurd [EMAIL PROTECTED] writes:
BH I am about to use the following tape type: HP C5708A DDS-3 Data
BH Cartridge, 24GB
I've used:
define tapetype DDS-3 {
comment DDS-3 DAT drive
length 12288 mbytes # 12 GB
filemark 0 kbytes
speed 850 kbytes
}
- J
MC == Mitch Collinsworth [EMAIL PROTECTED] writes:
MC So far no one has (to my knowledge) stepped forward and volunteered to
MC host the lists elsewhere, so this is what there is.
I have the facilities to do so. My list server machine isn't currently
very high-powered but this list has pretty
Does anyone have a tapetype for an Exabyte Mammoth2 drive (with 225m
tapes)? There doesn't seem to be one in the FAQ-O-Matic.
Thanks,
- J
53 matches
Mail list logo