Re: Amanda increment level?
On Tue, Apr 30, 2002 at 06:13:21AM +0300, Alexander Belik wrote: :) Will Amanda ask me for a new type when new dumpcycle begin? Or I must do it manualy? End of included message Alexander, Amanda uses a different model/concept than you are trying to fit on her. First, amanda expects each run of amdump will be on its own set of tapes. This means several tapes are in a rotation known as a tape cycle. My tape cycle is 18 tapes. Amanda will NOT add one dump to the end of a tape containing a previous dump. It will either refuse to use the tape (if your config says it is not time to reuse the tape) or it will overwrite the old data with a new amdump. Having a tapecycle of 1 means the next dump always destroys the ONLY dump you currently have. A definite NO-NO :) Second, amanda does not synchronize the levels of dump of each file system with each other one. Instead, each file system is scheduled independently with an aim of keeping the total size of each amdump run consistant. So you should excpect to see a mixture of levels in each amdump run. Third, to answer your specific question, no it will not ask. That sounds like you expect an interactive session. Amanda is designed to be run automatically on a scheduled basis (like my 2 AM nightly backup). And I'm not always here at 2 AM :) Instead, each amdump run generates a report, generally emailed to the admin, saying what tape it used and what tape it expects to have available the next amdump run. If that tape is not available for the next run, amanda complains (in the email report) and tries to do what she can. A common outcome is that the new amdump get written to a disk area known as the holding disk and can later be amflush'ed to the proper tape. HTH -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: Amanda increment level?
On Tue, 30 Apr 2002, Jon LaBadie wrote: :) Will Amanda ask me for a new type when new dumpcycle begin? Or I must do it manualy? First, amanda expects each run of amdump will be on its own set of tapes. This means several tapes are in a rotation known as a tape cycle. My tape cycle is 18 tapes. Ok! Thx! Amanda will NOT add one dump to the end of a tape containing a previous dump. It will either refuse to use the tape (if your config says it is not time to reuse the tape) or it will overwrite the old data with a new amdump. Having a tapecycle of 1 means the next dump always destroys the ONLY dump you currently have. A definite NO-NO :) So I must chage my types avery day (when amdump)? Second, amanda does not synchronize the levels of dump of each file system with each other one. Instead, each file system is scheduled independently with an aim of keeping the total size of each amdump run consistant. So you should excpect to see a mixture of levels in each amdump run. :) Thanks! I understand! Amanda decide when do a next level (depending on size only?) and avery dumpcycle runs level 0 dump! I'am right? Third, to answer your specific question, no it will not ask. That sounds like you expect an interactive session. Amanda is designed to be run automatically on a scheduled basis (like my 2 AM nightly backup). And I'm not always here at 2 AM :) Amnada after avery backup generate a report with line: --- The next tape Amanda expects to use is: AUTH2 So, I understand because tapecycle 1 tapes? Instead, each amdump run generates a report, generally emailed to the admin, saying what tape it used and what tape it expects to have available the next amdump run. If that tape is not available for the next run, amanda complains (in the email report) and tries to do what she can. A common outcome is that the new amdump get written to a disk area known as the holding disk and can later be amflush'ed to the proper tape. I know this! Thanks! -- [EMAIL PROTECTED]ICQ: 41776461 2:465/207@Fidonet Alexander Belik http://www.vnet.dn.ua[EMAIL PROTECTED]
Re: problem backing up large partition...
On Mon, 29 Apr 2002 at 9:11pm, Luke Miller wrote I am having some problems backing up a large partition (raid). I have a raid as a single partition with 56 GB of space. My tape drive is a DLT 20 GB drive with a tape library with 7 slots using mtx for the changer. I am running amanda version 2.4.2p2. I have set runtapes to 2 in amanda.conf hoping that I can use two tapes for my backups. Below is the backup report from this attempted backup. What am I missing here? The fact that amanda cannot span a partition across tapes (yet). Thus, no single disklist entry can be bigger than your tapelength. Or, what is the prefered method for doing a backup for my situation. Use GNUtar rather then dump (make sure you have at least version 1.13.19), and split the RAID into multiple disklist entries. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Backing up NT via smbclient - Problems
Hi, together, I've some problems backing up NT shares using smbclient/smbtar which is part of SAMBA. When the filename contains a german umlaut (i.e. ä, ö, ü, ß), an error occurs when amanda receives the files from the PC. Excerpt from mail report: ... /-- devwage700 //devwagwoce7621/backup1$ lev 1 STRANGE sendbackup: start [devwage7001://devwagwoce7621/backup1$ level 1] sendbackup: info BACKUP=/usr/local/bin/smbclient sendbackup: info RECOVER_CMD=/usr/local/bin/smbclient -f... - sendbackup: info end ? INFO: Debug class all level = 2 (pid 19398 from pid 19398) | added interface ip=10.210.8.170 bcast=10.210.11.255 nmask=255.255.252.0 [...] ? ERRDOS - ERRbadfile listing \Döring, Martin\* ? ERRDOS - ERRbadfile opening remote file \Ende, Jens\Eishotel\Mammut+Bär in einem Zimmer.jpg (\Ende, Jens\Eishotel\) [...] ? ERRDOS - ERRbadfile opening remote file \Projekte\Josua\Berichte\2000 - Josua Jahresabschluß.ppt (\Projekte\Josua\Berichte\) [...] ? ERRDOS - ERRbadfile listing \Schröder, Mark\* | tar: dumped 308 files and directories | Total bytes written: 293888 sendbackup: size 287 sendbackup: end \ Any idea? I tried to play around with the configuration on either the amanda and the samba side, but without success. My configuration: Amanda version 2.4.2p2 Solaris 8 Samba 2.2.2 Solaris 8 NT 4.0 SP6 shares Any help is appreciated. Thanks. With kind regards, Bjoern Briel Dr. Bjoern Briel Electronic Research Volkswagen AG, Brieffach 1776, 38436 Wolfsburg +49-5361-9-76239
Re: Backing up NT via smbclient - Problems
Hi, I've some problems backing up NT shares using smbclient/smbtar which is part of SAMBA. When the filename contains a german umlaut (i.e. ä, ö, ü, ß), an error occurs when amanda receives the files from the PC. Excerpt from mail report: [...] My configuration: Amanda version 2.4.2p2 @ Solaris 8 Samba 2.2.2 @ Solaris 8 NT 4.0 SP6 shares Use samba 2.2.3a, that removes this bug. It was introduced in 2.2.1 or 2.2.2 and solved in 2.2.3a. Make sure you have character set = ISO8859-1 or (-15) in your smb.conf. Regards, Christopher -- == Dipl.-Ing. Christopher Odenbach HNI Rechnerbetrieb [EMAIL PROTECTED] Tel.: +49 5251 60 6215 ==
Re: Qualstar 4222 Tape Library
On 26/04 2002 16:47 Frank Smith wrote: --On Friday, April 26, 2002 12:42:31 +0200 Toralf Lund [EMAIL PROTECTED] wrote: Has anyone used Amanda with the Qualstar 4222 tape library? We have a customer that want to use one for backup, and we probably have to help with the setup... We use a Qualstar 4480 (a bigger model in the same family) and it works fine with Amanda. What does your tape changer config look like? Also, I don't really know *anything* about tape libraries, but I need to provide some useful information to a customer who has one, but don't know what to do with it, so any kind of input is welcome. System is a small server running Red Hat Linux 7.2. - Toralf
OT: Backing up large raw partitions
Hi, I could need help in a somewhat unusual situation: A fellow PhD student is doing work time measurements with a proprietary digital video recoder based on Linux. His 120 GB disk is approaching full capacity and our only backup solution is the department server with its DAT DDS 4 tape drive (20 GB native) running Amanda for standard backups, of course. Without backup we have to process the data and delete it. I have not 'rooted' the box yet, but by inserting a graphics card I see Suse Linux 6.0 beeing used as an OS and that the large disk has no file system or partition table and might be used as a raw device (/dev/hdb). That's a clever way to circumvent nonexisting 2.0.36 Kernel handling of large files. That's ok as long as everything is done on the proprietary TV monitor interface, which has only local backup capabilities. But I don't know the data format, so for remote backups I need to split the raw device file in chunks of 19 GB and _reliably_ add them back. Not to think of incrementals. I can think of several ways to backup, but I've to get restores right the first time. Has someone _done_ something similar? I can think of several ways to split my backup: dd if=/dev/hdb of=/dev/rmt offset=... tar -Mcvf /dev/hdb /dev/rmt Thanks for proven receipts, Johannes Nieß P.S: I don't have a second 120 GB disk or equivalent disk storage space easily available. It's an unwieldy amount of data to handle :-( P.P.S: The vendor support has not answered my emails.
Re: Qualstar 4222 Tape Library
--On Tuesday, April 30, 2002 16:49:26 +0200 Toralf Lund [EMAIL PROTECTED] wrote: On 26/04 2002 16:47 Frank Smith wrote: --On Friday, April 26, 2002 12:42:31 +0200 Toralf Lund [EMAIL PROTECTED] wrote: Has anyone used Amanda with the Qualstar 4222 tape library? We have a customer that want to use one for backup, and we probably have to help with the setup... We use a Qualstar 4480 (a bigger model in the same family) and it works fine with Amanda. What does your tape changer config look like? I'm using the chg-qs-mtx script which Mark Holm modified from the chg-mtx script and posted to the list (search for Qualstar TLS 4212 in the archives and you'll find it, also get the fix in the followup). Relevant configurations are included below, modify to suit Before you get too involved in the changer script, make sure you can move tapes around using mtx from the command line, otherwise you will get very frustrated. Also, I don't really know *anything* about tape libraries, but I need to provide some useful information to a customer who has one, but don't know what to do with it, so any kind of input is welcome. System is a small server running Red Hat Linux 7.2. Libraries just let you further automate your backups, they don't do anything that you couldn't do sitting in front of a tape drive with a pile of tapes. Since there generally isn't someone next to the drive at all times, having a library lets you automate multi-tape backups and run backups nights, weekends, and holidays. They also keep your tapes from getting banged around so much. Good luck, Frank in amanda.conf= tpchanger chg-qs-mtx # tape-changer glue script tapedev /dev/nst0 # the no-rewind tape device to be used changerdev /dev/sg0 # the tape-changer device changerfile /usr/local/libexec/chg-qs-mtx.conf /usr/local/libexec/chg-qs-mtx.conf= # Number of the tape drive in changer as seen by MTX chgtapedev 0 # Barcode reader ? havereader 1 # First slot in Library to be used firstslot 1 # Last slot in Library to be used endslot 55 # Slot where the cleaning tape resides cleanslot 80 # How often do I clean (accesses of tape drive) cleancycle 1000 # Do we need to eject the tape: 1 - Yes, 0 - No needeject 1 = -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
what exactly is dtimeout?
Can someone tell me exactly what dtimeout in amanda.conf really is? The amanda(8) man page says: Amount of idle time per disk on a given client that a dumper running from within amdump will wait before it fails with a data timeout error. The docs directory has nothing further. I'm wondering what is meant by idle time. Is this maximum wall clock time dumper will allow for a dump to finish? Or does idle really mean idle and it's more like amount of time dumper will wait without having received any new data from sendbackup? I have a largeish dump that's getting the [data timeout] error, but it's not stalling. It appears it's just not being allowed to finish. If dtimeout has the former meaning then increasing it should solve this. If it has the latter meaning then I need to look elsewhere. -Mitch
Re: OT: Backing up large raw partitions
On Tue, 2002-04-30 at 09:51, Johannes Niess wrote: Hi, I could need help in a somewhat unusual situation: A fellow PhD student is doing work time measurements with a proprietary digital video recoder based on Linux. His 120 GB disk is approaching full capacity and our only backup solution is the department server with its DAT DDS 4 tape drive (20 GB native) running Amanda for standard backups, of course. Without backup we have to process the data and delete it ... Has someone _done_ something similar? I can think of several ways to split my backup: dd if=/dev/hdb of=/dev/rmt offset=... tar -Mcvf /dev/hdb /dev/rmt If it were me, I would do a gnutar-wrapper or dump-wrapper type solution (i.e have amanda call your script instead of dump or tar), and call it 12 10G pseudo-partitions, named something like /dev/hdb/0 through /dev/hdb/11. Use a shell script and dd skip=`expr 10 '*' $slice`G bs=1M count=10240 to actually generate the dump image. Then to restore it you would do something like: for slice in 0 1 2 3 4 5 6 7 8 9 10 11 do amadmin config find host /dev/hdb/$slice\$ echo mount tape, press enter when ready read line amrestore hostname -p host /dev/hdb/$slice date | rsh host dd seek=`expr 10 '*' $slice`G of=/dev/hdb done
Turning down messages
Is there any way to turn down the level of error messages or warnings displayed from tar? I'm getting these: /-- orca / lev 3 STRANGE sendbackup: start [orca:/ level 3] sendbackup: info BACKUP=/bin/tar sendbackup: info RECOVER_CMD=/bin/gzip -dc |/bin/tar -f... - sendbackup: info COMPRESS_SUFFIX=.gz sendbackup: info end ? gtar: ./dev/log: socket ignored | Total bytes written: 3983360 (3.8MB, 486kB/s) sendbackup: size 3890 sendbackup: end \ Which are making my logs and email large (some have a bunch of sockets that are ignored). I don't want to lose valuable information, though. Thanks for any help you can provide. -- Terri Eads Systems Administrator [EMAIL PROTECTED] Research Applications Program (303) 497-8425 National Center for Atmospheric Research
Re: Turning down messages
On Tue, 30 Apr 2002 at 10:45am, Terri Eads wrote Is there any way to turn down the level of error messages or warnings displayed from tar? I'm getting these: Strings which are considered normal are listed as DMP_NORMAL in client-src/sendbackup-gnutar.c. But... /-- orca / lev 3 STRANGE sendbackup: start [orca:/ level 3] sendbackup: info BACKUP=/bin/tar sendbackup: info RECOVER_CMD=/bin/gzip -dc |/bin/tar -f... - sendbackup: info COMPRESS_SUFFIX=.gz sendbackup: info end ? gtar: ./dev/log: socket ignored I see that message in there in version 2.4.2p2. What version of amanda are you using? Upgrading to 2.4.2p2 should fix this. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Turning down messages
Hi Terri - You should setup the gtar exclude list so you can specify files you want to ignore. -doug On Tue, 30 Apr 2002, Terri Eads wrote: Is there any way to turn down the level of error messages or warnings displayed from tar? I'm getting these: /-- orca / lev 3 STRANGE sendbackup: start [orca:/ level 3] sendbackup: info BACKUP=/bin/tar sendbackup: info RECOVER_CMD=/bin/gzip -dc |/bin/tar -f... - sendbackup: info COMPRESS_SUFFIX=.gz sendbackup: info end ? gtar: ./dev/log: socket ignored | Total bytes written: 3983360 (3.8MB, 486kB/s) sendbackup: size 3890 sendbackup: end \ Which are making my logs and email large (some have a bunch of sockets that are ignored). I don't want to lose valuable information, though. Thanks for any help you can provide. -- ~~ Doug Silver Network Manager Urchin Corporation http://www.urchin.com ~~
problems with GNUtar and --sparse
I am having some problems with GNU tar and amanda. I am running amanda 2.4.2p2 with gnutar 1.13.19 (I have also tried 1.13.25) on Solaris 8 (sparc). The command line that I am running is: /usr/local/bin/tar --create --file /tmp/conf.tar --directory /export/raid/conf --one-file-system --sparse --listed-incremental /private/amanda/var/amanda/gnutar-lists/nfs-1c-bvtn_export_raid_conf_5.new --totals . Replace /tmp/conf.tar for - when it is run by amanda. When I run this is produces a tar file that is about 5 megs in size and contains 2556 entries. If I run the same command line but without the --sparse I get a tar file that is about 23 megs and has 9535 enties. So, what is the --sparse doing here that is causing it not to backup all the files? What is the best way around this? I would hack the code to remove --sparse but I want to have an understanding of what it is doing first. Thanks, Luke * * Luke Miller Unix System Administrator * * Integra Telecom 503-748-4549 * *
Re: problems with GNUtar and --sparse
Look at the GNU tar manual (not the man page) for a good explanation of what sparse files are and how --sparse handles them: http://www.gnu.org/manual/tar-1.12/html_node/tar_121.html Frank --On Tuesday, April 30, 2002 15:49:24 -0700 Luke Miller [EMAIL PROTECTED] wrote: I am having some problems with GNU tar and amanda. I am running amanda 2.4.2p2 with gnutar 1.13.19 (I have also tried 1.13.25) on Solaris 8 (sparc). The command line that I am running is: /usr/local/bin/tar --create --file /tmp/conf.tar --directory /export/raid/conf --one-file-system --sparse --listed-incremental /private/amanda/var/amanda/gnutar-lists/nfs-1c-bvtn_export_raid_conf_5.new --totals . Replace /tmp/conf.tar for - when it is run by amanda. When I run this is produces a tar file that is about 5 megs in size and contains 2556 entries. If I run the same command line but without the --sparse I get a tar file that is about 23 megs and has 9535 enties. So, what is the --sparse doing here that is causing it not to backup all the files? What is the best way around this? I would hack the code to remove --sparse but I want to have an understanding of what it is doing first. Thanks, Luke * * Luke Miller Unix System Administrator * * Integra Telecom 503-748-4549 * * -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: problems with GNUtar and --sparse
That's helpful but I don't think it explains what I am seeing. When I have the --sparse option turned on, tar does not backup all the files that it should. There is at least one directory where it doesn't backup ANY files at all in it. Thanks, Luke Look at the GNU tar manual (not the man page) for a good explanation of what sparse files are and how --sparse handles them: http://www.gnu.org/manual/tar-1.12/html_node/tar_121.html Frank --On Tuesday, April 30, 2002 15:49:24 -0700 Luke Miller millerlu@integraonli ne.com wrote: I am having some problems with GNU tar and amanda. I am running amanda 2.4 .2p2 with gnutar 1.13.19 (I have also tried 1.13.25) on Solaris 8 (sparc). The command line that I am running is: /usr/local/bin/tar --create --file /tmp/conf.tar --directory /export/raid/c onf --one-file-system --sparse --listed-incremental /private/amanda/var/amanda/ gnutar-lists/nfs-1c-bvtn_export_raid_conf_5.new --totals . Replace /tmp/conf.tar for - when it is run by amanda. When I run this is produces a tar file that is about 5 megs in size and con tains 2556 entries. If I run the same command line but without the --sparse I get a tar file th at is about 23 megs and has 9535 enties. So, what is the --sparse doing here that is causing it not to backup all th e files? What is the best way around this? I would hack the code to remove --sparse but I want to have an understanding of what it is doing first. Thanks, Luke * * Luke Miller Unix System Administrator * * Integra Telecom 503-748-4549 * * -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: problems with GNUtar and --sparse
--On Tuesday, April 30, 2002 16:24:35 -0700 Luke Miller [EMAIL PROTECTED] wrote: That's helpful but I don't think it explains what I am seeing. When I have the --sparse option turned on, tar does not backup all the files that it should. There is at least one directory where it doesn't backup ANY files at all in it. Sorry, I missed the part about the number of files being less, I just thought the total size was less than expected. I have no idea how --sparse could affect the nuber of files. Good luck, Frank Thanks, Luke Look at the GNU tar manual (not the man page) for a good explanation of what sparse files are and how --sparse handles them: http://www.gnu.org/manual/tar-1.12/html_node/tar_121.html Frank --On Tuesday, April 30, 2002 15:49:24 -0700 Luke Miller millerlu@integraonli ne.com wrote: I am having some problems with GNU tar and amanda. I am running amanda 2.4 .2p2 with gnutar 1.13.19 (I have also tried 1.13.25) on Solaris 8 (sparc). The command line that I am running is: /usr/local/bin/tar --create --file /tmp/conf.tar --directory /export/raid/c onf --one-file-system --sparse --listed-incremental /private/amanda/var/amanda/ gnutar-lists/nfs-1c-bvtn_export_raid_conf_5.new --totals . Replace /tmp/conf.tar for - when it is run by amanda. When I run this is produces a tar file that is about 5 megs in size and con tains 2556 entries. If I run the same command line but without the --sparse I get a tar file th at is about 23 megs and has 9535 enties. So, what is the --sparse doing here that is causing it not to backup all th e files? What is the best way around this? I would hack the code to remove --sparse but I want to have an understanding of what it is doing first. Thanks, Luke * * Luke Miller Unix System Administrator * * Integra Telecom 503-748-4549 * * -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501 -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Invalid Argument with amlabel
I searched the list archives and read all the documentation and I cannot find this error anywhere. $ amlabel gcs DailySet1 insert tape into slot 0 and press return labeling tape in slot 0 (/dev/nst0): rewinding, reading label, not an amanda tape rewinding, writing label DailySet1 amlabel: writing label: Invalid argument Notice the Invalid argument. I have a Seagate Python 04687 SCSI tape backup drive. I am using Amanda 2.4.2p2 which I compiled myself on the same machine, which is Linux 2.4.13 and glibc 2.2.2. The backup user is backup. I am using nst0 as my tape device and backup has permission to read/write: # ls -l /dev/nst0 crw-rw1 root disk 9, 128 Apr 14 2001 /dev/nst0 # cat /etc/passwd | grep backup backup:x:417:6:Amanda Backup:/root/amanda-data:/bin/false # cat /etc/group | grep backup disk:x:6:root,backup Here is some of my amanda.conf: tpchanger chg-manual tapedev /dev/nst0 changerfile /root/amanda-data/changer-status changerdev /dev/null tapetype DAT labelstr ^DailySet[0-9][0-9]*$ infofile /root/amanda-data logdir /var/log/amanda indexdir /root/amanda-data tapelist /root/amanda-data/tapelist define tapetype DAT { comment Archive Python 04687-XXX length 4000 mbytes filemark 100 kbytes speed 800 kbytes } amcheck returns no errors except that the current loaded tape has no label: $ amcheck gcs Amanda Tape Server Host Check - Holding disk /mnt/spare/var/amanda/: 1731365 KB disk space available, that's plenty insert tape into slot 1 and press return amcheck-server: slot 1: not an amanda tape I obviously need to label my tapes. Anyone have any ideas? Thanks, Cory Visi