unsubscribe amanda-users
unsubscribe amanda-users
Re: tape error
Hi, R. Zoontjens schrieb: Hi, I have a customer who is running our unix server with amanda for 2 months now. Every night after a dump we get an email with the dump-statistics. The dumps are all full dumps and we used 5 tapes for it (which we labeled). dums are from Mo till Fr (5 days). No problems so far. Last night we got a tape error: *** A TAPE ERROR OCCURRED: [writing label: Read-only file system] * Does this mean that the tape is write-protected? AFAIK yes, shurely someone set the physical writeprotection of the cartridge. * Is amanda trying to label the tape? Why: we already labeled it amanda rewrites the label every time it overwites a tape. that is the first action she does on the tape. * Is this critical? yes, more or less. No data was written to the tape as it was writeprotected. amanda will probably insist on using this tape in the next run again. Christoph I tried a google search and list achives but coun't find an answer Met vriendelijke groet / with kind regards, R.J. Zoontjens Radécom B.V. [EMAIL PROTECTED] www.radecom.nl __ / __ \ _/ /_/ _ ___ / /_/ / __ `/ __ / _ \/ ___/ __ \/ __ `__ \ / _, _/ /_/ / /_/ / __/ /__/ /_/ / / / / / / /_/ |_|\__,_/\__,_/\___/\___/\/_/ /_/ /_/ Dit bericht kan vertrouwelijke informatie bevatten. Indien u niet de geadresseerde van dit bericht bent, verzoeken wij u dit bericht te vernietigen zonder van de inhoud kennis te nemen en de inhoud ervan niet te gebruiken, niet te kopiëren en niet onder derden te verspreiden. This message may contain information which is privileged or confidential. If you are not the named addressee of this message please destroy it without reading, using, copying or disclosing its contents to any other person.
Re: concerns
On Mon, 4 Nov 2002 at 6:33pm, Galen Johnson wrote After further study of the restore issues I've had with amanda (mostly tar related) I have just today run into what I feel is another problem (I suspect is a combination of tar and amanda in this case). I am concerned with the incremental backups. Amanda seems to treat incrementals as differentials in the way it labels them for later recovery when used in conjunction with tar (I have no idea how it works with dump). The way I understand incrementals is that and incremental is just the files changed since the last incremental or full. A differential is the files changed since the last full backup (from what I've seen this is how amanda treats the incremental whereby it only increments the level by 1 if it meets certain criteria (which I'm fairly certain are definable in amanda.conf)). Amanda with tar works just as with dump, including with regards to incrementals. An incremental is anything less than a full, a.k.a. a level 0. There are multiple levels of incrementals. Any level backs up everything that has changed since the last backup of a lower level. I.e. a level 1 backs up everything changed since the last level 0. A level 2 backs up everything changed since the last level 1. Etc etc etc. I hope I've made clear what I'm trying to point out. The incremental as amanda deals with it is really being treated as a differential in regards to a restore. Can anyone recommend any suggestions to make this behavior a bit more along the accepted norms? I was thinking of setting the default amanda.conf bumpsize of 20 Mb to 1 Kb to see if it will set the incrementals to give truer incrementals. Any thoughts on this matter are greatly appreciated. Well, you could do that, and get higher levels, but why would you want to? If all your incrementals are level 1s, then you'll only ever need two tapes to do a full restore -- the last level 0, and the latest level 1. The more levels of incrementals you use, the more tapes you'll need to restore from. If you've got the tape space, why not use it? In short, it's a feature meant to save you time at restore time and protect you against the possibility of bad tapes. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
How to run a single dump by hand?
Last night's run gave the following: *** A TAPE ERROR OCCURRED: [[writing file: Input/output error]]. Some dumps may have been left in the holding disk. Run amflush to flush them to tape. The next tape Amanda expects to use is: DailySet1-02. FAILURE AND STRANGE DUMP SUMMARY: amundsen /export/amundsen3 lev 0 FAILED [missing result for /export/amundsen3 in amundsen response] orca /export/local lev 0 FAILED [out of tape] amundsen failed because of a configuration problem, orca because it ran out of tape. Before running amflush to write the orca dump to tape, I'd like to dump the amundsen directory so they both can be written to tape. How hard is this? Thanks! - Orion
Re: How to run a single dump by hand?
On Tue, Nov 05, 2002 at 08:56:21AM -0700, Orion Poplawski wrote: Last night's run gave the following: *** A TAPE ERROR OCCURRED: [[writing file: Input/output error]]. Some dumps may have been left in the holding disk. Run amflush to flush them to tape. The next tape Amanda expects to use is: DailySet1-02. FAILURE AND STRANGE DUMP SUMMARY: amundsen /export/amundsen3 lev 0 FAILED [missing result for /export/amundsen3 in amundsen response] orca /export/local lev 0 FAILED [out of tape] amundsen failed because of a configuration problem, orca because it ran out of tape. Before running amflush to write the orca dump to tape, I'd like to dump the amundsen directory so they both can be written to tape. How hard is this? easy... if you have 2.4.3 Remove the tape from the drive amdump config amundsen /export/amundsen3 Put a tape on the drive amflush config Jean-Louis -- Jean-Louis Martineau email: [EMAIL PROTECTED] Departement IRO, Universite de Montreal C.P. 6128, Succ. CENTRE-VILLETel: (514) 343-6111 ext. 3529 Montreal, Canada, H3C 3J7Fax: (514) 343-5834
Re: How to run a single dump by hand?
Orion Poplawski wrote: Last night's run gave the following: *** A TAPE ERROR OCCURRED: [[writing file: Input/output error]]. Some dumps may have been left in the holding disk. Run amflush to flush them to tape. The next tape Amanda expects to use is: DailySet1-02. FAILURE AND STRANGE DUMP SUMMARY: amundsen /export/amundsen3 lev 0 FAILED [missing result for /export/amundsen3 in amundsen response] orca /export/local lev 0 FAILED [out of tape] amundsen failed because of a configuration problem, orca because it ran out of tape. Before running amflush to write the orca dump to tape, I'd like to dump the amundsen directory so they both can be written to tape. How hard is this? Thanks! - Orion It's been my experience that as long as amanda sees something in the holding disk any subsequent dump go directly to holding disk as well. You can then either select a day at a time or have amanda try to dump all holding to a single disk (I tend to do this even though it can be a bad idea). =G=
GNU tar file problem with long filename
We're using amanda and GNU tar (1.13.25) to backup a new Netapp to a Solaris 8 Sun. The first pilot users moved to the Netapp went fine, and then amanda started acting up. The symptom was tar was grinding on one of the CPUs and amanda was marking the host as not responding. The host happens to also be the amanda server, but I don't think that is relevant. truss indicated that there were NO system calls being made by the process. Captured the amanda ps process being used and duplicated it on the command line, adding the --verbose flag. What I've traced it back to, is one of the users I migrated to the Netapp had a very long file name in his .kde http cache ./xxx/.kde/share/cache/http/r/www.redhat.com_docs_manuals_linux_RHL-7.3-Manual_getting-started-guide_figs_disks_xcdr1.png_5a13234a Obviously nothing he tried to create. Now the pathname is 131 chars long, and the basename of the file is 101 chars. By renaming the file tar went along happily. I did some web and Amanda archive searches looking for a reference to this problem and came up with nothing on point. Of course I always second guess myself on the search criteria used. Did I miss a known issue with gnu tar? And I thought it odd that the error condition would be to sit and spin on a CPU. Has anyone else seen similar behavior? Thanks, Murf /usr/local/gnu/bin/amanda-tar --version tar (GNU tar) 1.13.25 Copyright (C) 2001 Free Software Foundation, Inc. This program comes with NO WARRANTY, to the extent permitted by law. You may redistribute it under the terms of the GNU General Public License; see the file named COPYING for details. Written by John Gilmore and Jay Fenlason. root 7554 45.8 0.5 2200 1936 ?R 18:00:01 811:11 /usr/local/gnu/bin/amanda-tar --create --file /dev/null --directory /moria/vol/vol0/Research --one-file-system --listed-incremental /usr/local/amanda-2.4.3/gnutar-lists/castle_moria_vol_vol0_Research_0.new --sparse --ignore-failed-read --totals --exclude-from /tmp/amanda/sendsize._moria_vol_vol0_Research.20021031180001.exclude . -- John Murphy, better known as Erin's Kevin's dadhttp://remarque.org/~jam [EMAIL PROTECTED] One one-trillionith of a surprise: picoboo 345 Scarborough Road millihellen: The beauty needed to launch 1 ship Briarcliff Manor, NY 10510 ** My opinions do not reflect those of my employer **
Re: How to run a single dump by hand?
On Tuesday 05 November 2002 09:27 am, Jean-Louis Martineau wrote: On Tue, Nov 05, 2002 at 08:56:21AM -0700, Orion Poplawski wrote: Before running amflush to write the orca dump to tape, I'd like to dump the amundsen directory so they both can be written to tape. How hard is this? easy... if you have 2.4.3 Remove the tape from the drive amdump config amundsen /export/amundsen3 Put a tape on the drive amflush config This finally prompted me to whip up some 2.4.3 rpms and put them on the server. This worked great once I setup the reserve space properly. Thanks! - Orion
Re: GNU tar file problem with long filename
I wonder if you create a long file by that name with touch then try to tar and untar it if it will report an error? --- Murf [EMAIL PROTECTED] wrote: We're using amanda and GNU tar (1.13.25) to backup a new Netapp to a Solaris 8 Sun. The first pilot users moved to the Netapp went fine, and then amanda started acting up. The symptom was tar was grinding on one of the CPUs and amanda was marking the host as not responding. The host happens to also be the amanda server, but I don't think that is relevant. truss indicated that there were NO system calls being made by the process. Captured the amanda ps process being used and duplicated it on the command line, adding the --verbose flag. What I've traced it back to, is one of the users I migrated to the Netapp had a very long file name in his .kde http cache ./xxx/.kde/share/cache/http/r/www.redhat.com_docs_manuals_linux_RHL-7.3-Manual_getting-started-guide_figs_disks_xcdr1.png_5a13234a Obviously nothing he tried to create. Now the pathname is 131 chars long, and the basename of the file is 101 chars. By renaming the file tar went along happily. I did some web and Amanda archive searches looking for a reference to this problem and came up with nothing on point. Of course I always second guess myself on the search criteria used. Did I miss a known issue with gnu tar? And I thought it odd that the error condition would be to sit and spin on a CPU. Has anyone else seen similar behavior? Thanks, Murf /usr/local/gnu/bin/amanda-tar --version tar (GNU tar) 1.13.25 Copyright (C) 2001 Free Software Foundation, Inc. This program comes with NO WARRANTY, to the extent permitted by law. You may redistribute it under the terms of the GNU General Public License; see the file named COPYING for details. Written by John Gilmore and Jay Fenlason. root 7554 45.8 0.5 2200 1936 ?R 18:00:01 811:11 /usr/local/gnu/bin/amanda-tar --create --file /dev/null --directory /moria/vol/vol0/Research --one-file-system --listed-incremental /usr/local/amanda-2.4.3/gnutar-lists/castle_moria_vol_vol0_Research_0.new --sparse --ignore-failed-read --totals --exclude-from /tmp/amanda/sendsize._moria_vol_vol0_Research.20021031180001.exclude . -- John Murphy, better known as Erin's Kevin's dad http://remarque.org/~jam [EMAIL PROTECTED] One one-trillionith of a surprise: picoboo 345 Scarborough Road millihellen: The beauty needed to launch 1 ship Briarcliff Manor, NY 10510 ** My opinions do not reflect those of my employer ** __ Do you Yahoo!? HotJobs - Search new jobs daily now http://hotjobs.yahoo.com/
amdump: Negative estimates!
Hello all,
I just installed amanda Amanda 2.4.3 on a machine called thebe2/Solaris 8
This machine a tape drive connected to it, and the amlabel worked fine
on it.
Now when I try to dump a test filesystem, /tmp, or even /, I get disk
off line error, details follow.
Any idea what is it that I forgot?
Thanks.
Mohamed~
# cat disklist
thebe2 /tmp { # the line break here is mandatory
root-tar # copy properties of root-tar
compress server fast # but change the compression mode
} # spindle and interface omitted here
I did set GNUTAR to the right binary, which is :
# tar --version
tar (GNU tar) 1.13.25
amcheck reports no problems. then when I try amdump, I get this in
amdumplog:
SETTING UP FOR ESTIMATES...
planner: time 0.007: setting up estimates for thebe2:/tmp
thebe2:/tmp overdue 11996 days for level 0
setup_estimate: thebe2:/tmp: command 0, options:
last_level -1 next_level0 -11996 level_days 0
getting estimates 0 (0) -1 (-1) -1 (-1)
...
ANALYZING ESTIMATES...
planner: FAILED thebe2 /tmp 20021104 0 [disk /tmp offline on thebe2?]
INITIAL SCHEDULE (size 80):
Re: GNU tar file problem with long filename
On Tue, Nov 05, 2002 at 12:17:20PM -0800, Jerry wrote: I wonder if you create a long file by that name with touch then try to tar and untar it if it will report an error? --- Murf [EMAIL PROTECTED] wrote: We're using amanda and GNU tar (1.13.25) to backup a new Netapp to a Solaris 8 Sun. The first pilot users moved to the Netapp went fine, and then amanda started acting up. The symptom was tar was grinding on one of the CPUs ... . What I've traced it back to, is one of the users I migrated to the Netapp had a very long file name in his .kde http cache Obviously nothing he tried to create. Now the pathname is 131 chars long, and the basename of the file is 101 chars. By renaming the file tar went along happily. ... Has anyone else seen similar behavior? I've a Solaris 8, x86 version with two gnutars, 1.13 that comes on the Companion CD and a self-compiled 1.13.25 in addition to the standard tar. I created a directory with a set of 15Kb files, each with names 5, 15, 25, ... long out to 305 chars. The OS prevented creation of anything beyond 255. The standard /usr/bin/tar bombed out at anything over 100 chars. The two gnutar's both handled creation of a tarball of this directory and of recovery from either's tarball. I did not check if a long pathname to the directory would make a difference. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
index question
Is it proper for amanda to not see the indexes for disks that have been removed from the disklist but are still on the current tapes? I recently stopped backing up a server to keep from getting a bunch of failed messages in my reports. Now when I run amrecover and setdisk to that backup it tells me that there are no indexes and yet I can look in my var/daily/index/hostname directory and the indexes are there. Is this normal? =G=
Re: concerns
Jon LaBadie has been corresponding with me offlist in regards to this topic. I will summarize when I am confident I have found the culprit. (whether it be Amanda, tar or the actual disk being backed up)... =G=
Re: index question
On Tue, 5 Nov 2002 at 5:08pm, Galen Johnson wrote Is it proper for amanda to not see the indexes for disks that have been removed from the disklist but are still on the current tapes? I recently stopped backing up a server to keep from getting a bunch of failed messages in my reports. Now when I run amrecover and setdisk to that backup it tells me that there are no indexes and yet I can look in my var/daily/index/hostname directory and the indexes are there. Is this normal? I believe that amrecover does the equivalent of a amadmin CONFIG find $HOST $DISK to look for disks to recover, and this relies on the disklist. So, yes it's normal. You can either put the entry temporarily back in the disklist or get the stuff via amrestore. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: index question
On Tue, Nov 05, 2002 at 05:23:59PM -0500, Joshua Baker-LePain wrote: On Tue, 5 Nov 2002 at 5:08pm, Galen Johnson wrote Is it proper for amanda to not see the indexes for disks that have been removed from the disklist but are still on the current tapes? I recently stopped backing up a server to keep from getting a bunch of failed messages in my reports. Now when I run amrecover and setdisk to that backup it tells me that there are no indexes and yet I can look in my var/daily/index/hostname directory and the indexes are there. Is this normal? I believe that amrecover does the equivalent of a amadmin CONFIG find $HOST $DISK to look for disks to recover, and this relies on the disklist. So, yes it's normal. You can either put the entry temporarily back in the disklist or get the stuff via amrestore. Joshua is right. I comment the disklist entry when I remove a disk. I uncomment it if I need to restore it. But I don't like that. A solution could be to add a new option in a dumptype active [yes|no] alive [yes|no] backup [yes|no] dobackup [yes|no] That say if amanda must do backup of this disk, but the disk will be known to all other amanda command. You will need to change the dumptype of a disk when it is removed. Another solution could be to add it directly in the disklist file, as a prefix before the hostname. DEAD host.dead.com /disk/dead What do you think should be the best solution? Jean-Louis -- Jean-Louis Martineau email: [EMAIL PROTECTED] Departement IRO, Universite de Montreal C.P. 6128, Succ. CENTRE-VILLETel: (514) 343-6111 ext. 3529 Montreal, Canada, H3C 3J7Fax: (514) 343-5834
Re: index question
Jean-Louis Martineau wrote: snip Joshua is right. I comment the disklist entry when I remove a disk. I uncomment it if I need to restore it. But I don't like that. A solution could be to add a new option in a dumptype active [yes|no] alive [yes|no] backup [yes|no] dobackup [yes|no] That say if amanda must do backup of this disk, but the disk will be known to all other amanda command. You will need to change the dumptype of a disk when it is removed. Another solution could be to add it directly in the disklist file, as a prefix before the hostname. DEAD host.dead.com /disk/dead What do you think should be the best solution? Jean-Louis I personally think editting just one file would be simpler hence just editting the disklist entry since the dumptypes are generally used by multiple disks. Even better would be some logic that amanda is able to associate date with a tape number and be intelligent enough to decide for itself if the disk should be included in all the necessary commands regardless of whether it lives in the disklist or not...this should be fairly easy to do with the indexing (he says without even looking). Of course after further thought, you would just have to assign a different definition to the dumptype list and assign that to the dead disk...I'd say this would come down to the developers deciding which would be easier to implement. I'd like it either way... =G=
Re: tar/ not missing any new directories
At 10:56 AM 10/14/2002 +0200, Toralf Lund wrote:
With tar, and some sort of a guarantee that no individual file will
exceed the tape capacity, this can be done by breaking the disklist
entries up into subdirs,
Yes, that's what I'm doing. The problem with this is that something easily
gets left out as new directories are created.
I'm late to this conversation, but I don't think it is possible to
leave out any new directories with a scheme like this:
define dumptype diskA-TheRest {
comp-user-tar #or other local globally defined TAR dumptype
exclude list /diskA/diskA.exclude
}
on client,/diskA/diskA.exclude contains:
fred
sally
sam
tom
on server, disklist:
client.fqdn /diskA diskA-TheRest #excludes fred,sally,sam,tom
client.fqdn /diskA/fred comp-user-tar
client.fqdn /diskA/sally comp-user-tar
client.fqdn /diskA/sam comp-user-tar
client.fqdn /diskA/tom comp-user-tar
(a) when remembered, add a new directory by
editting disklist (explicitly
add deb and client:/diskA/diskA.exclude explicitly exclude deb
(b) when forgotten, new directory is already included in the
diskA-TheRest because you haven't explicitly excluded it.
Deb Baddorf
---
Deb Baddorf [EMAIL PROTECTED] 840-2289
You can't help getting older, but you don't have to get old.
- George Burns IXOYE
Flaþ!..Flaþ!..Flaþ!.. Dünyanýn Ýlk Astroloji ve Gizli Ýlimler Portalý Açýldý...
Dünya'nn lk Astroloji ve Gizli limler Portalwww.astromerkez.com Dünya'nn ilk astroloji portal hizmete girdi. Haftalk maç yorumlar, Bu Haftaki Fener-Galata Derbisinin sonucu Gündemdeki konular, Deprem, Saddam ve Sava, Seçimler, AB ve Türkiye'nin gelecei , R.Tayyip Erdoan'n ve Deniz Baykal' n astroloji haritalar ve karakter analizleri, kiiye özel ücretsiz detayl günlük fal hizmeti, Tüm burçlarn hiçbir yerde bulamayacanz detayl analizleri, Tüm gizemli limler, Vefk, Remil, Ebced, Havvas,Ledün... ifal dualar, Duann tesirinin bilimsel açklamas ve yaplan aratrma ve deneyler, Manyetizma, Durugörü, Astral Seyahat, Meditasyon, Ruhu Bedenden Ayrma.. Ve Daha Pek Çok Konu Tüm Dünya'da ilk defa ve sadece www.astromerkez.com da
