amanda hangs on using holding disk
Hello, I use redhat7 with amanda 2.4.2p2 rpm. I tryed to backup several servers by using holding disk. Without holding Disk the backup works ok, but with holding disk, the backup hangs after an hour or two. Is this a well known failure in this amanda version? How can I fix it? Regards Stephan Helas
Avoiding incremental backups (how?)
I have been trying to make an archival tape configuration for some time now. I have tried numerous suggestions, but nothing seems to be effective in stopping Amanda from doing incremental backups. I am using Amanda 2.4.3b4.
The point of this configuration is to run once every 4 weeks and always do a full backup to tape. There are two tapes in the rotation and one is aways off-site.
My amanda.conf:
dumpcycle 0
runspercycle 2
tapecycle 2 tapes
define dumptype comp-root-full {
record no
index yes
comment Root partitions with compression
compress client fast
priority low
skip-incr yes
}
define dumptype comp-user-full {
record no
index yes
comment Non-root partitions on reasonably fast machines
compress client fast
priority medium
skip-incr yes
}
define dumptype comp-user-full-gnutar {
record no
index yes
program GNUTAR
comment user partitions dumped with tar and compression
compress server fast
maxdumps 4
exclude ./RECYCLER
skip-incr yes
}
My disklist:
server02.ctsda1comp-root-full1
server02.ctsda2comp-user-full1
server02.ctsdd1comp-user-full2
server02.ct//SERVER01/BACKUP-Ccomp-user-full-gnutar
server02.ct//SERVER01/BACKUP-Ecomp-user-full-gnutar
If you reply to this, please reply to me directly in addition to the list, as I am no longer subscribed.
Thanks,
Cory Visi
Re: What device does what on st0 with a Tecmar Travan-NS20?
Wayne Johnson wrote: I just bought a Tecmar Travan NS20 IDE tape drive. Took me a while to figure out that I should use the IDE-SCSI shim to get it to work. One question. There are 4 modes for the driver, tied to device names /dev/st0, /dev/st0l, /dev/st0m, and /dev/st0a. Anyone know what the different modes are? I would guess that one is with compression on. Anyone know which does what? I've started running typetype, but if the research I found is true, it'll take 5 hours for each device type. Thanks. From /usr/src/linux/Documentation/devices.txt: /dev/st0 First SCSI tape, mode 0 /dev/st0l First SCSI tape, mode 1 /dev/st0m First SCSI tape, mode 2 /dev/st0a First SCSI tape, mode 3 My understanding from having had to research this in the past: /dev/st0 - normal rewinding tape device /dev/st0l - normal rewinding, block density as defined in sys/mtio.h or by /etc/stinit /dev/st0m - normal rewinding compressed device /dev/st0a - normal rewinding, drive buffer as defined in sys/mtio.h or by /etc/stinit You can get this info (hopefully) by doing a 'man st' and reading through sys/mtio.h. You should also have the sam non-rewinding tape devices (/dev/nst0,...). On most systems I've only ever seen /dev/st0 and /dev/nst0. The other devices would have had to havee been created manually using mknod. On some systems the no-rewind device is /dev/st0n. Note: On some systems the different modes can be defined in the file /dev/stinit. Note: I'm not absolutely certain about the l and a components. =G=
More than one dump per server
Hi All I was wondering if someone can tel me how to configure amanda to run more than one dump per server? Thanks
Backup Scenario
Hi everyone, If I want to backup 2 partitions of the localhost in 2 tapes, is that possible by doing something like that? dumpcycle 0 days # for full backups runspercycle 5 # backup weekdays only ??? tapecycle 5 tapes runtapes 2 #Do I have to define this?? #tpchanger chg-manual localhost /var always-full # I want this to tape 1 localhost /home always-full # This to tape 2 Also how should I label the tapes, can I give them any name?? Can someone help please?? Thanks Steve
Re: amanda hangs on using holding disk
On Thu, Oct 31, 2002 at 09:58:33AM +0100, Helas, Stephan wrote: Hello, I use redhat7 with amanda 2.4.2p2 rpm. I tryed to backup several servers by using holding disk. Without holding Disk the backup works ok, but with holding disk, the backup hangs after an hour or two. Is this a well known failure in this amanda version? How can I fix it? A guess, your system has a file size limit of 2GB and some dump files exceed that size. Second, you do not have chunksize set in amanda.conf. Set chunksize to say 1 GB to avoid the problem. Amanda will then break up the dump file into 1GB size chunks on the holding disk. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: A couple of questions
Hi 1.- If a single dump image (one disc partition) is longer than a tape, amanda can't handle this. If multiple dump images dont fit in a tape you can setup a changer-device=manual which will be a script you can write to send an email, a terminal prompt, windows popup, sing a song or whatever to notify that the operator has to replace the tape. I have not used it so I can be wrong. 2.- We do so also. We have one config for Daily tapes and other for Weekly/offsite/fulldumps. In your case: Daily - dumpcycle1 week runspercycle 5 (monday to friday) runtapes 1 (If you use 1 tape each day) tapecycle6 (Or more. You at least need 1 tape extra to not destroy the first valid tape in the cycle in the event of a hardware faillure during the run) If you need, say 2 tapes a day, then runtapes=2, tapecycle=12 Weekly - dumpcycle0 (full dumps) runspercycle 1 runtapes 1 (If your weekly dumps fits in 1 tape) tapecycle2 (1 in site, 1 offsite. Or more) record no If you need, say 3 tapes for the weekly backup, then runtapes=3, tapecycle=6 (list, please let me know if I'm wrong!) Jordi On Thu, 31 Oct 2002 10:03:39 0500 Steve Simeonidis [EMAIL PROTECTED] wrote. I need to install a backup software on a few servers and I was told that amanda is really good. What I'm not sure about: 1/ Some of the data we backup do not fit on a single tape How do I handle that if we haven't got a changer?? 2/ Usually we do full backups only. We have 7 tapes, 5 weekday and 2 that are used as weekly tapes that are send to another location. How can I configure that with amanda? I'm a bit confused with dumpcycle-runspercycle-tapecycle-runtapes Thanks Steve
Re: What device does what on st0 with a Tecmar Travan-NS20?
On Fri, 25 Oct 2002 at 5:08pm, Wayne Johnson wrote One question. There are 4 modes for the driver, tied to device names /dev/st0, /dev/st0l, /dev/st0m, and /dev/st0a. Anyone know what the different modes are? I would guess that one is with compression on. Anyone know which does what? /dev/nst0 is the device you want. N is for non-rewinding, which amanda requires. Also, do the Travan drives even do hardware compression? ISTR that they don't... -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: A couple of questions
On Thu, 31 Oct 2002 at 10:03am, Steve Simeonidis wrote 1/ Some of the data we backup do not fit on a single tape How do I handle that if we haven't got a changer?? 2/ Usually we do full backups only. We have 7 tapes, 5 weekday and 2 that are used as weekly tapes that are send to another location. How can I configure that with amanda? I'm a bit confused with dumpcycle-runspercycle-tapecycle-runtapes If you'll look through the archives, you'll see this is a very common beginner's question. Typically, forcing amanda to do what you want doesn't work as well as letting it do what it wants -- it's good at it's job. As for your questions, if you have partitions that are bigger than a tape, you can split them up into multiple disklist entries if you specify GNUTAR as your dump program. You say you do full backups only, but some data sets are bigger than a tape -- how can you do weeklies then? In your situation, I'd do this: For your nightly backups, define a Daily configuration. dumpcycle 1 week, runspercycle 5, and tapecycle 10 (you'll need to buy more tapes -- having tapecycle=runspercycle is Bad, as one failed backup can mean you destroy your only backup of some systems). For the weeklies, dumpcycle 0, tapecycle 2 (or however many you need to fit all your data). Good luck. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: A couple of questions
Thanks for your reply. If I go according to your examples I need 2 config files. When do you do your weekly backup? Is the Daily backup only incremental?? I thought (may be I'm wrong) it does full backup autotically depending on the dumpcycle size. I can't find anything on the changer-device script anywhere, where can I place that script, where do I call it from. Thanks again On Thursday 31 October 2002 17:53, jordivi wrote: Hi 1.- If a single dump image (one disc partition) is longer than a tape, amanda can't handle this. If multiple dump images dont fit in a tape you can setup a changer-device=manual which will be a script you can write to send an email, a terminal prompt, windows popup, sing a song or whatever to notify that the operator has to replace the tape. I have not used it so I can be wrong. 2.- We do so also. We have one config for Daily tapes and other for Weekly/offsite/fulldumps. In your case: Daily - dumpcycle1 week runspercycle 5 (monday to friday) runtapes 1 (If you use 1 tape each day) tapecycle6 (Or more. You at least need 1 tape extra to not destroy the first valid tape in the cycle in the event of a hardware faillure during the run) If you need, say 2 tapes a day, then runtapes=2, tapecycle=12 Weekly - dumpcycle0 (full dumps) runspercycle 1 runtapes 1 (If your weekly dumps fits in 1 tape) tapecycle2 (1 in site, 1 offsite. Or more) record no If you need, say 3 tapes for the weekly backup, then runtapes=3, tapecycle=6 (list, please let me know if I'm wrong!) Jordi On Thu, 31 Oct 2002 10:03:39 0500 Steve Simeonidis [EMAIL PROTECTED] wrote. I need to install a backup software on a few servers and I was told that amanda is really good. What I'm not sure about: 1/ Some of the data we backup do not fit on a single tape How do I handle that if we haven't got a changer?? 2/ Usually we do full backups only. We have 7 tapes, 5 weekday and 2 that are used as weekly tapes that are send to another location. How can I configure that with amanda? I'm a bit confused with dumpcycle-runspercycle-tapecycle-runtapes Thanks Steve
Re: A couple of questions
Unfortunately the documentation on amanda is not up to scratch. I've been trying for a few days now to get this thing to work. 8-( I've gone through the archives but it is very hard sometimes to find exactly what you are looking for. When I changed the amanda configuration file to use runtapes 2 tpchanger chg-manual and try to label the tape I get the following error. any ideas? [rootzeus tmp]# su amanda -c amlabel -f DailySet1 DMP01 bash: /root/.bashrc: Permission denied labeling tape in slot /usr/lib/amanda/chg-manual: (/usr/adm/amanda/changer-status-clean: No such file or directory): rewinding amlabel: no tape online Also I've noticed that some people use an extra parameter in their disklist configuration file, what is it used for? I can't find any docs. eg localhost /var always-full 0 What is that 0 value for, some people use 1. Thanks On Thursday 31 October 2002 18:04, Joshua Baker-LePain wrote: On Thu, 31 Oct 2002 at 10:03am, Steve Simeonidis wrote 1/ Some of the data we backup do not fit on a single tape How do I handle that if we haven't got a changer?? 2/ Usually we do full backups only. We have 7 tapes, 5 weekday and 2 that are used as weekly tapes that are send to another location. How can I configure that with amanda? I'm a bit confused with dumpcycle-runspercycle-tapecycle-runtapes If you'll look through the archives, you'll see this is a very common beginner's question. Typically, forcing amanda to do what you want doesn't work as well as letting it do what it wants -- it's good at it's job. As for your questions, if you have partitions that are bigger than a tape, you can split them up into multiple disklist entries if you specify GNUTAR as your dump program. You say you do full backups only, but some data sets are bigger than a tape -- how can you do weeklies then? In your situation, I'd do this: For your nightly backups, define a Daily configuration. dumpcycle 1 week, runspercycle 5, and tapecycle 10 (you'll need to buy more tapes -- having tapecycle=runspercycle is Bad, as one failed backup can mean you destroy your only backup of some systems). For the weeklies, dumpcycle 0, tapecycle 2 (or however many you need to fit all your data). Good luck.
Timing out can't find out why.
The machine is up and the amanda server and the client can ping eachother. All running Solaris 2.6. On same net/subnet as other clients that don't time out. Any suggestions??? I upped the etimeout value from 500 to 5500 -Original Message- From: [EMAIL PROTECTED] [mailto:bin;AmServer-JP.genuity2.net] Sent: Thursday, October 31, 2002 1:33 AM To: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: AOL-JAPAN-REMOTE AMANDA PROBLEM: FIX BEFORE RUN, IF POSSIBLE Amanda Tape Server Host Check - /hold: 2680048 KB disk space available, that's plenty. amcheck-server: slot 13: date 20021030 label JAPAN-28 (active tape) amcheck-server: slot 14: date 20021016 label JAPAN-29 (exact label match) NOTE: skipping tape-writable test. Tape JAPAN-29 label ok. Server check took 175.929 seconds. Amanda Backup Client Hosts Check WARNING: Japan-Collect1b: selfcheck request timed out. Host down? Client check: 9 hosts checked in 30.183 seconds, 1 problems found. (brought to you by Amanda 2.4.0)
Can't determine disk and mount point from $CWD
Hi. I'm trying to run amrecover but I get this message up above the amrecover prompt:- Can't determine disk and mount point from $CWD everything else up to this point looks ok:- #amrecover Bacup AMRECOVER Version 2.4.2p2. Contacting server on Neli ... 220 Neli AMANDA index server (2.4.2p2) ready. 200 Access OK Setting restore date to today (2002-10-31) 200 Working date set to 2002-10-31. 200 Config set to Bacup. 200 Dump host set to Neli. Can't determine disk and mount point from $CWD amrecover I have no idea how to solve this! Any help or advice would be greatly appreciated. Thank you
Re: Can't determine disk and mount point from $CWD
On Thu, 31 Oct 2002 at 2:27pm, [EMAIL PROTECTED] wrote Hi. I'm trying to run amrecover but I get this message up above the amrecover prompt:- Can't determine disk and mount point from $CWD [jlb@chaos jlb]$ man amrecover . . . setdisk diskname [ mountpoint ] Specifies which disk to consider (default: the disk holding the working directory where amrecover is started). It can only be set after the host is set with sethost. Diskname is the device name speci fied in the amanda.conf or disklist configuration file. The disk must be local to the host. If mountpoint is not specified, all pathnames will be relative to the (unknown) mount point instead of full pathnames. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
WARNING: dilmom: selfcheck request timed out. Host down?
Hi, We have had Amanda running for a while now - just recently we re-booted our server and are getting the following when AMCHECK is run - have I missed something. WARNING: dilmom: selfcheck request timed out. Host down? Client check: 1 host checked in 30.005 seconds, 1 problem found (brought to you by Amanda 2.4.3b3) Thanks Kevin Passey
Re: Avoiding incremental backups (how?)
runspercycle needs to be 1 to do what you want.
Frank
--On Thursday, October 31, 2002 04:13:11 -0500 Cory Visi [EMAIL PROTECTED] wrote:
I have been trying to make an archival tape configuration for some time
now. I have tried numerous suggestions, but nothing seems to be
effective in stopping Amanda from doing incremental backups. I am using
Amanda 2.4.3b4.
The point of this configuration is to run once every 4 weeks and always
do a full backup to tape. There are two tapes in the rotation and one is
aways off-site.
My amanda.conf:
dumpcycle 0
runspercycle 2
tapecycle 2 tapes
define dumptype comp-root-full {
record no
index yes
comment Root partitions with compression
compress client fast
priority low
skip-incr yes
}
define dumptype comp-user-full {
record no
index yes
comment Non-root partitions on reasonably fast machines
compress client fast
priority medium
skip-incr yes
}
define dumptype comp-user-full-gnutar {
record no
index yes
program GNUTAR
comment user partitions dumped with tar and compression
compress server fast
maxdumps 4
exclude ./RECYCLER
skip-incr yes
}
My disklist:
server02.ctsda1comp-root-full
1
server02.ctsda2comp-user-full
1
server02.ctsdd1comp-user-full
2
server02.ct//SERVER01/BACKUP-C
comp-user-full-gnutar
server02.ct//SERVER01/BACKUP-E
comp-user-full-gnutar
If you reply to this, please reply to me directly in addition to the
list, as I am no longer subscribed.
Thanks,
Cory Visi
--
Frank Smith[EMAIL PROTECTED]
Systems Administrator Voice: 512-374-4673
Hoover's Online Fax: 512-374-4501
Re: More than one dump per server
--On Thursday, October 31, 2002 13:33:46 +0200 Du-Wayne Rood [EMAIL PROTECTED] wrote: Hi All I was wondering if someone can tel me how to configure amanda to run more than one dump per server? Thanks If you mean back up more than one filesystem, just add another line in the disklist with the additional path. If you mean back up the same data twice (such as for offsite storage), then you need a separate config and need to be aware of the issues involved. Frank -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: Backup Scenario
--On Thursday, October 31, 2002 22:07:48 +0500 Steve Simeonidis [EMAIL PROTECTED] wrote: Hi everyone, If I want to backup 2 partitions of the localhost in 2 tapes, is that possible by doing something like that? dumpcycle 0 days # for full backups runspercycle 5 # backup weekdays only ??? tapecycle 5 tapes runtapes 2 # Do I have to define this?? # tpchanger chg-manual localhost /var always-full # I want this to tape 1 localhost /home always-full # This to tape 2 Also how should I label the tapes, can I give them any name?? Can someone help please?? Thanks Steve If you want them to go to separate tapes, you need separate configs. Frank -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: tru64 5.1 2.4.2p2
More stuff: I changed libtool to use ksh, but it still fails with ../libtool: print: not found . print does work in a test script in interactive shell. Don't know if it can't find print or is it complaining about the parameter passed to print ( $echo. ) I tried 2.4.3 on 4.0G - it too fails with that error. The libtools generated in 2.4.3 2.4.2p2 differ. I copied the 2.4.2p2 libtool to 2.4.3 it built fine. Installed, amcheck'd the backups ran fine last night. -- toby bluhm philips medical systems, mr development, cleveland, ohio [EMAIL PROTECTED] 440-483-5323 Jean-Louis Martineau [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] 10/28/2002 05:19 PM To: Tobias Bluhm/CLE/MS/PHILIPS@AMEC cc: [EMAIL PROTECTED] Subject:Re: tru64 5.1 2.4.2p2 Classification: On Mon, Oct 28, 2002 at 03:47:44PM -0500, [EMAIL PROTECTED] wrote: I'm using /bin/sh First line libtool: #! /sbin/sh Echo in libtool: # An echo program that does not interpret backslashes. echo=print -r Could you change the first line of libtool to use ksh instead of sh? What's the relationship between the libtool above ./configure --disable-libtool which is how I ran configure ( and user, group,etc. ) Don't know, that may be a bug in autoconf/automake. Jean-Louis -- toby bluhm philips medical systems, mr development, cleveland, ohio [EMAIL PROTECTED] 440-483-5323 Jean-Louis Martineau [EMAIL PROTECTED] 10/28/2002 01:58 PM To: Tobias Bluhm/CLE/MS/PHILIPS@AMEC cc: [EMAIL PROTECTED] Subject:Re: tru64 5.1 2.4.2p2 Classification: Hi tobias, Which shell are you using? What's the first line of libtool? What's the value of 'echo' in the libtool program? Jean-Louis On Mon, Oct 28, 2002 at 09:39:41AM -0500, [EMAIL PROTECTED] wrote: I tried 2.4.3 get a libtool error ( as jfkoenig suggested I would! ) /sbin/sh ../libtool --mode=link cc -g -o genversion genversion.o versuff.o alloc.o clock.o debug.o error.o util.o file.o -lm -ltermcap ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found ../libtool: print: not found rm -f version.c ./genversion version.c sh: ./genversion: not found *** Exit 1 Stop. *** Exit 1 Stop. I did however, copy the configure'd amanda 2.4.2p2 directory from one of our Tru64 4.x boxes, ran make install. The backups ran over the weekend without a problem. -- toby bluhm philips medical systems, mr development, cleveland, ohio [EMAIL PROTECTED] 440-483-5323 [EMAIL PROTECTED] Sent by: [EMAIL PROTECTED] 10/24/2002 12:58 PM To: [EMAIL PROTECTED] cc: (bcc: Tobias Bluhm/CLE/MS/PHILIPS) Subject:tru64 5.1 2.4.2p2 Classification: amanda 2.4.2p2 on tru64 5.1 - make stops on error: cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src-g -c alloc.c cc: Warning: amanda.h, line 946: In this declaration, parameter 1 has a different type than specified in an earlier declaration of this function. (mismatparam) extern int ruserok P((const char *rhost, int suser, ---^ cc: Error: amanda.h, line 946: In this declaration, the type of ruserok is not compatible with the type of a previous declaration of ruserok at line number 705 in file /usr/include/unistd.h. (notcompat) extern int ruserok P((const char *rhost, int suser, ---^ *** Exit 1 Stop. *** Exit 1 Stop. Found a message in the archives that points to ftp://gandalf.cc.purdue.edu/pub/amanda/configure.jfkoenig as a possible solution, but ftp on gandalf.cc.purdue.edu does not seem to be up. Would that file help me? Does anyone have a copy to send me? Thanks -- toby bluhm philips medical systems, mr development, cleveland, ohio [EMAIL PROTECTED] 440-483-5323 -- 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 -- 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
timeout
The machine is up and the amanda server and the client can ping eachother. All running Solaris 2.6. On same net/subnet as other clients that don't time out. Any suggestions??? I upped the etimeout value from 500 to 5500 Amanda Tape Server Host Check - /hold: 2680048 KB disk space available, that's plenty. amcheck-server: slot 13: date 20021030 label JAPAN-28 (active tape) amcheck-server: slot 14: date 20021016 label JAPAN-29 (exact label match) NOTE: skipping tape-writable test. Tape JAPAN-29 label ok. Server check took 175.929 seconds. Amanda Backup Client Hosts Check WARNING: Japan-Collect1b: selfcheck request timed out. Host down? Client check: 9 hosts checked in 30.183 seconds, 1 problems found. (brought to you by Amanda 2.4.0)
Re: timeout
On Thu, 31 Oct 2002 at 11:38am, Siobhan Gowen wrote The machine is up and the amanda server and the client can ping eachother. All running Solaris 2.6. On same net/subnet as other clients that don't time out. Any suggestions??? I upped the etimeout value from 500 to 5500 This has nothing to do with etimeout. Amanda Backup Client Hosts Check WARNING: Japan-Collect1b: selfcheck request timed out. Host down? Client check: 9 hosts checked in 30.183 seconds, 1 problems found. Is there anything in /tmp/amanda/amandad*debug on Japan-Collect1b? Make sure the debug file you look at is time-stamped at the time you ran amcheck. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
forcing amanda to not go into degraded mode
in my amanda.conf file I have the line: reserve 10 # percent which as I read the comments, means that amanda will still do full backups, if they'll fit in the other 90% of the 80GB disk I'm using as a 'holding' disk. however, when the dump runs, I get the ever-popular: lev 0 FAILED [dumps too big, but cannot incremental dump skip-incr disk] error for a number of filesystems that didn't make it onto tape. :( (even tho the disk being backed up is 500MB, and should happily fit in the 40+GB of space on the holding disk that ought to be available in any case) What I'm trying for here, is to have the backup dump as much to tape as it can, and then leave the rest on disk, so I can run amflush the next morning when I come in. (we don't have a tape changer, and backups need to be automated, so chg-manual isn't an option as I understand its usage). is there some setting that I'm missing, to tell amanda don't go into degraded mode when dumping to disk after a tape error? Carl Soderstrom. -- Systems Administrator Real-Time Enterprises www.real-time.com
RE: timeout
This log exists on all the machines except the one that's failing. I just got this dumped on me so I'm not very familiar with amanda. I can't seem to find it ever completing successfully. Although since it's in Japan our day is their night so a find is scanning the holding disk. Are there relative logs on the amanda backup server I could look at? -Original Message- From: [EMAIL PROTECTED] [mailto:owner-amanda-users;amanda.org]On Behalf Of Joshua Baker-LePain Sent: Thursday, October 31, 2002 11:58 AM To: Siobhan Gowen Cc: Alex Specogna; [EMAIL PROTECTED] Subject: Re: timeout On Thu, 31 Oct 2002 at 11:38am, Siobhan Gowen wrote The machine is up and the amanda server and the client can ping eachother. All running Solaris 2.6. On same net/subnet as other clients that don't time out. Any suggestions??? I upped the etimeout value from 500 to 5500 This has nothing to do with etimeout. Amanda Backup Client Hosts Check WARNING: Japan-Collect1b: selfcheck request timed out. Host down? Client check: 9 hosts checked in 30.183 seconds, 1 problems found. Is there anything in /tmp/amanda/amandad*debug on Japan-Collect1b? Make sure the debug file you look at is time-stamped at the time you ran amcheck. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: forcing amanda to not go into degraded mode
--On Thursday, October 31, 2002 11:18:34 -0600 Carl Wilhelm Soderstrom [EMAIL PROTECTED] wrote: in my amanda.conf file I have the line: reserve 10 # percent which as I read the comments, means that amanda will still do full backups, if they'll fit in the other 90% of the 80GB disk I'm using as a 'holding' disk. You've got it backwards, the 10% you reserved is used for fulls, the other 90% is used for incrementals. Frank however, when the dump runs, I get the ever-popular: lev 0 FAILED [dumps too big, but cannot incremental dump skip-incr disk] error for a number of filesystems that didn't make it onto tape. :( (even tho the disk being backed up is 500MB, and should happily fit in the 40+GB of space on the holding disk that ought to be available in any case) What I'm trying for here, is to have the backup dump as much to tape as it can, and then leave the rest on disk, so I can run amflush the next morning when I come in. (we don't have a tape changer, and backups need to be automated, so chg-manual isn't an option as I understand its usage). is there some setting that I'm missing, to tell amanda don't go into degraded mode when dumping to disk after a tape error? Carl Soderstrom. -- Systems Administrator Real-Time Enterprises www.real-time.com -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: timeout
My guess is that amandad is not running. Check the inetd.conf and services to make sure they are correct. Siobhan Gowen wrote: This log exists on all the machines except the one that's failing. I just got this dumped on me so I'm not very familiar with amanda. I can't seem to find it ever completing successfully. Although since it's in Japan our day is their night so a find is scanning the holding disk. Are there relative logs on the amanda backup server I could look at? -Original Message- From: [EMAIL PROTECTED] [mailto:owner-amanda-users;amanda.org]On Behalf Of Joshua Baker-LePain Sent: Thursday, October 31, 2002 11:58 AM To: Siobhan Gowen Cc: Alex Specogna; [EMAIL PROTECTED] Subject: Re: timeout On Thu, 31 Oct 2002 at 11:38am, Siobhan Gowen wrote The machine is up and the amanda server and the client can ping eachother. All running Solaris 2.6. On same net/subnet as other clients that don't time out. Any suggestions??? I upped the etimeout value from 500 to 5500 This has nothing to do with etimeout. Amanda Backup Client Hosts Check WARNING: Japan-Collect1b: selfcheck request timed out. Host down? Client check: 9 hosts checked in 30.183 seconds, 1 problems found. Is there anything in /tmp/amanda/amandad*debug on Japan-Collect1b? Make sure the debug file you look at is time-stamped at the time you ran amcheck. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University -- Paul T. Root CCSA, CCSE, CCNA Qwest Communications PAG: +1 (877) 693-7155 600 Stinson Blvd, Flr 1S WRK: +1 (612) 664-3385 Minneapolis, MN 55413 FAX: +1 (612) 664-4778
Re: forcing amanda to not go into degraded mode
You've got it backwards, the 10% you reserved is used for fulls, the other 90% is used for incrementals. hmmm, the comments read completely the opposite of that, but I'll give it a try. thanks much! Carl Soderstrom. -- Systems Administrator Real-Time Enterprises www.real-time.com
sun l25 library
is anyone using amanda with the sun L25 tape library, or the quantum ATL M1500 library as they are one in the same? i would imagine that it should work fine using a combination of the sgen driver, mtx, and chg-zd-mtx. i'm interested in finding out if anyone has experiences, good/bad with the device before i purchase. thanks. -- Darin Perusich Unix Systems Administrator Cognigen Corp. [EMAIL PROTECTED]
Re: forcing amanda to not go into degraded mode
From: Carl Wilhelm Soderstrom [EMAIL PROTECTED] however, when the dump runs, I get the ever-popular: lev 0 FAILED [dumps too big, but cannot incremental dump skip-incr disk] I don't think Amanda is telling you that the level 0 won't fit on the holding disk. It is telling you that it won't fit on the tape. It isn't even considering putting it on the holding disk, because it knows it can't fit it on the tape. If you want to have Amanda backup more than will fit on the tape and leave the rest for a manual flush, as you describe, you could try a tapetype that claims the tape is much larger than it actually is. The other option is the manual changer configuration. I haven't used this, but my understanding is that you tell Amanda that you have 2 (or more tapes), and that manual intervention is required to go to subsequent tapes. Amanda will fill the first tape for you overnight, and in the morning you change tapes, tell Amanda, and it will use the second tape. No need for a manual flush in this scenario. Doug -- Doug Meredith Skyridge Systems Inc. (506) 854-7997 www.skyridge.com
Re: tru64 5.1 2.4.2p2
On Thu, Oct 31, 2002 at 11:29:31AM -0500, [EMAIL PROTECTED] wrote: More stuff: I changed libtool to use ksh, but it still fails with ../libtool: print: not found . print does work in a test script in interactive shell. Don't know if it can't find print or is it complaining about the parameter passed to print ( $echo. ) I tried 2.4.3 on 4.0G - it too fails with that error. The libtools generated in 2.4.3 2.4.2p2 differ. I copied the 2.4.2p2 libtool to 2.4.3 it built fine. Installed, amcheck'd the backups ran fine last night. I know, it's a bug in the newer libtool. Could you try this: untar amanda ./configure edit config.status change - s%@LIBTOOL@%$(SHELL) $(top_builddir)/libtool% to - s%@LIBTOOL@%/bin/ksh $(top_builddir)/libtool% make 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
Tapetype for a Sony SDX-500C
I have searched the FAQ and the archives but I can find a typetype for a Sony SDX-500C which is an AIT-2 tape drive. If anyone can help I would appreciate it. = Kathy Ange Virginia Department of Agriculture Consumer Services Information Systems (804) 786-1340 Voice Mail (804) 786-2110 FAX __ Do you Yahoo!? HotJobs - Search new jobs daily now http://hotjobs.yahoo.com/
Re: Tapetype for a Sony SDX-500C
Try this:
define tapetype AIT-2 {
comment SDX-500C
length 54538 mbytes
filemark 1541 kbytes
speed 2920 kps
}
It seems a little off but works.
Make sure compression is turned off on the drive or you will
only get about 44GB on a tape.
Frank
--On Thursday, October 31, 2002 12:00:59 -0800 Kathy Ange [EMAIL PROTECTED] wrote:
I have searched the FAQ and the archives but I can
find a typetype for a Sony SDX-500C which is an AIT-2
tape drive. If anyone can help I would appreciate it.
=
Kathy Ange
Virginia Department of Agriculture Consumer Services
Information Systems
(804) 786-1340 Voice Mail
(804) 786-2110 FAX
__
Do you Yahoo!?
HotJobs - Search new jobs daily now
http://hotjobs.yahoo.com/
--
Frank Smith[EMAIL PROTECTED]
Systems Administrator Voice: 512-374-4673
Hoover's Online Fax: 512-374-4501
Re: Tapetype for a Sony SDX-500C
Thanks I will give it a try
--- Frank Smith [EMAIL PROTECTED] wrote:
Try this:
define tapetype AIT-2 {
comment SDX-500C
length 54538 mbytes
filemark 1541 kbytes
speed 2920 kps
}
It seems a little off but works.
Make sure compression is turned off on the drive or
you will
only get about 44GB on a tape.
Frank
--On Thursday, October 31, 2002 12:00:59 -0800 Kathy
Ange [EMAIL PROTECTED] wrote:
I have searched the FAQ and the archives but I can
find a typetype for a Sony SDX-500C which is an
AIT-2
tape drive. If anyone can help I would appreciate
it.
=
Kathy Ange
Virginia Department of Agriculture Consumer
Services
Information Systems
(804) 786-1340 Voice Mail
(804) 786-2110 FAX
__
Do you Yahoo!?
HotJobs - Search new jobs daily now
http://hotjobs.yahoo.com/
--
Frank Smith
[EMAIL PROTECTED]
Systems Administrator
Voice: 512-374-4673
Hoover's Online
Fax: 512-374-4501
__
Do you Yahoo!?
HotJobs - Search new jobs daily now
http://hotjobs.yahoo.com/
Re: forcing amanda to not go into degraded mode
I don't think Amanda is telling you that the level 0 won't fit on the holding disk. It is telling you that it won't fit on the tape. no, in that case it usually gives the 'dumps too big for tape' error message. If you want to have Amanda backup more than will fit on the tape and leave the rest for a manual flush, as you describe, you could try a tapetype that claims the tape is much larger than it actually is. I already turned my supposed tape length up 10GB larger than the actual tape; and some of these partitions that are failing, only have 500MB of data. The other option is the manual changer configuration. I haven't used this, but my understanding is that you tell Amanda that you have 2 (or more tapes), and that manual intervention is required to go to subsequent tapes. Amanda will fill the first tape for you overnight, and in the morning you change tapes, tell Amanda, and it will use the second tape. No need for a manual flush in this scenario. the faq-o-matic seems to indicate that the chg-manual configuration only works when started interactively (so you have a console with which to send a signal to amanda that there's a new tape in the drive). I know there's ways to delay backups (I use them for some machines already), so even if I started backups at 6pm, machines could be delayed until the early morning; but I don't trust myself or the tape monkey to remember (or be available) to start a backup every night. Carl Soderstrom. -- Systems Administrator Real-Time Enterprises www.real-time.com
Linux backup of IRIX failing
I'm trying to run amanda on a Linux server and have it backup a mix of clients. I'm trying to test out an IRIX client, and the backup is failing with: FAILURE AND STRANGE DUMP SUMMARY: orca /export/acmap1 lev 0 FAILED [disk /export/acmap1 offline on orca?] orca /export/local lev 0 FAILED [disk /export/local offline on orca?] orca /export/contrib lev 0 FAILED [disk /export/contrib offline on orca?] amcheck reports no problems. amdump.1 log shows: setting up estimates for orca:/export/contrib orca:/export/contrib overdue 11991 days for level 0 setup_estimate: orca:/export/contrib: command 0, options: last_level -1 next_level0 -11991 level_days 0 getting estimates 0 (0) -1 (-1) -1 (-1) setting up estimates for orca:/export/local orca:/export/local overdue 11991 days for level 0 setup_estimate: orca:/export/local: command 0, options: last_level -1 next_level0 -11991 level_days 0 getting estimates 0 (0) -1 (-1) -1 (-1) setting up estimates for orca:/export/acmap1 orca:/export/acmap1 overdue 11991 days for level 0 setup_estimate: orca:/export/acmap1: command 0, options: last_level -1 next_level0 -11991 level_days 0 getting estimates 0 (0) -1 (-1) -1 (-1) Do the amanda installations on the servers and clients need to reference the same paths? The IRIX installation is from SGI's freeware site and puts files in different locations. Is there any other debugging I can run to see what's failing on the client? - Orion
Re: forcing amanda to not go into degraded mode
--On Thursday, October 31, 2002 16:14:28 -0600 Carl Wilhelm Soderstrom [EMAIL PROTECTED] wrote: I don't think Amanda is telling you that the level 0 won't fit on the holding disk. It is telling you that it won't fit on the tape. no, in that case it usually gives the 'dumps too big for tape' error message. If you want to have Amanda backup more than will fit on the tape and leave the rest for a manual flush, as you describe, you could try a tapetype that claims the tape is much larger than it actually is. I already turned my supposed tape length up 10GB larger than the actual tape; and some of these partitions that are failing, only have 500MB of data. The other option is the manual changer configuration. I haven't used this, but my understanding is that you tell Amanda that you have 2 (or more tapes), and that manual intervention is required to go to subsequent tapes. Amanda will fill the first tape for you overnight, and in the morning you change tapes, tell Amanda, and it will use the second tape. No need for a manual flush in this scenario. the faq-o-matic seems to indicate that the chg-manual configuration only works when started interactively (so you have a console with which to send a signal to amanda that there's a new tape in the drive). I know there's ways to delay backups (I use them for some machines already), so even if I started backups at 6pm, machines could be delayed until the early morning; but I don't trust myself or the tape monkey to remember (or be available) to start a backup every night. If you have enough holding disk, don't configure a changer and specify an absurdly large tapetype. Then Amanda will do all the backups required and give you an EOT error in your daily report. Then run amflush in the morning as many times as required (each time you hit EOT it will error out and email you, then change tapes and run it again. This way your backups will all occur at the scheduled time and you can flush during the day when you're available to change tapes. The emails will remind you to do the tape changing. Frank Carl Soderstrom. -- Systems Administrator Real-Time Enterprises www.real-time.com -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: forcing amanda to not go into degraded mode
If you have enough holding disk, don't configure a changer and specify an absurdly large tapetype. tried that. 40GB should be enough to hold everything, but I'll turn it up to 70GB (size of the holding disk). Then Amanda will do all the backups required and give you an EOT error in your daily report. yep. do that already. thanks much for the advice! :) Carl Soderstrom. -- Systems Administrator Real-Time Enterprises www.real-time.com
Re: Linux backup of IRIX failing
On Thu, 31 Oct 2002 at 3:30pm, Orion Poplawski wrote I'm trying to run amanda on a Linux server and have it backup a mix of clients. I'm trying to test out an IRIX client, and the backup is failing with: FAILURE AND STRANGE DUMP SUMMARY: orca /export/acmap1 lev 0 FAILED [disk /export/acmap1 offline on orca?] orca /export/local lev 0 FAILED [disk /export/local offline on orca?] orca /export/contrib lev 0 FAILED [disk /export/contrib offline on orca?] Are those all partitions? If not, are you specifying GNUTAR as your dump program? What's in /tmp/amanda/sendsize*debug on orca? Do the amanda installations on the servers and clients need to reference the same paths? The IRIX installation is from SGI's freeware site and puts files in different locations. Hrm. You may want to try compiling and installing on orca -- that way you know what goes where and from who. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Linux backup of IRIX failing
On Thursday 31 October 2002 04:07 pm, Joshua Baker-LePain wrote: On Thu, 31 Oct 2002 at 3:30pm, Orion Poplawski wrote I'm trying to run amanda on a Linux server and have it backup a mix of clients. I'm trying to test out an IRIX client, and the backup is failing with: FAILURE AND STRANGE DUMP SUMMARY: orca /export/acmap1 lev 0 FAILED [disk /export/acmap1 offline on orca?] orca /export/local lev 0 FAILED [disk /export/local offline on orca?] orca /export/contrib lev 0 FAILED [disk /export/contrib offline on orca?] Are those all partitions? If not, are you specifying GNUTAR as your dump program? What's in /tmp/amanda/sendsize*debug on orca? Forgot about /tmp/amanda. sendsize shows: calculating for amname '/export/acmap1', dirname '/export/acmap1' sendsize: getting size via gnutar for /export/acmap1 level 0 sendsize: missing exclude list file /usr/local/lib/amanda/exclude.gtar discarded sendsize: spawning /usr/freeware/libexec/runtar in pipeline sendsize: argument list: /usr/freeware/bin/tar --create --file /dev/null --directory /export/acmap1 --one-file-system --listed-incremental /var/amanda/gnutar-lists/orca_export_acmap1_0.new --sparse --ignore-failed-read --totals . runtar: error [must be invoked by root] Looks like amanda was compiled with --with-user=root? I'll try running as root on the client... Do the amanda installations on the servers and clients need to reference the same paths? The IRIX installation is from SGI's freeware site and puts files in different locations. Hrm. You may want to try compiling and installing on orca -- that way you know what goes where and from who. I initially tried that, but was running into strange bugs (/usr/bin/install appears to be broken). We'll see... Thanks! - Orion
chg-manual
amanda-2.4.2p2 Can someone explain to me how the chg-manual works? I want to use more than one tape with runtape 2 When I enable the chg-manual option and try to use something like amlabel the system complains. [rootzeus tmp]# su amanda -c amlabel -f DailySet1 DMP01 slot 0 bash: /root/.bashrc: Permission denied insert tape into slot 0 and press return labeling tape in slot /usr/lib/amanda/chg-manual: (/usr/adm/amanda/changer-status-clean: No such file or directory): rewinding amlabel: no tape online There is a tape in there, and it ejects it when I run the amlabel command, I put the tape back in press enter but it comes back with an error. Also the /usr/adm/amanda/changer-status-clean does not exist on my system. Can someone give me more info?? Thanks Steve
