incremental backup without tape
I'd like to perform incremental backup of my servers leaving it on holding-disk Is it possible to do? How can I do it? Thanks a lot -- Alessandro Chiauzzi System Engineer Dialogos3 srl begin:vcard n:Chiauzzi;Alessandro tel;cell:0338/4435939 tel;work:02/70005949 x-mozilla-html:FALSE org:Dialogos3;Network Administration adr:;;via Carlo Poma 18;Milano;Italy;20124; version:2.1 email;internet:[EMAIL PROTECTED] x-mozilla-cpt:;0 fn:Alessandro Chiauzzi end:vcard
RV: amreport
What version of Amanda are you using? i am using amanda 2.4.2 Did you get a core file in that directory or /tmp/amanda? If so, what do you get when you run a debugger on amreport and the core file and enter the "where" command, e.g.: $ gdb /path/to/amreport /path/to/core gdb where ... gdb quit I include the output of the gdb. I don't understand anything. Amreport is normally run for you as part of amdump. Why are you running it by hand (not that that's bad, I'm just curious)? ok, i thought it was as you are saying, but i espected amanda to mailme with the results, but it did'd happen, and it didn't show the statistic either. So i changed my opinion and thought i should run it myself. John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
RV: amreport
Hello: I've discovered that one of the errors is generated because there isn't a tapelist file, but i thought it was generated by amanda itself. Is that right??? i have the line: tapelist "/usr/local/etc/amanda/Diario/tapelist" , that is, in the same directory as amanda.conf thanks mo What version of Amanda are you using? i am using amanda 2.4.2 Did you get a core file in that directory or /tmp/amanda? If so, what do you get when you run a debugger on amreport and the core file and enter the "where" command, e.g.: $ gdb /path/to/amreport /path/to/core gdb where ... gdb quit I include the output of the gdb. I don't understand anything. Amreport is normally run for you as part of amdump. Why are you running it by hand (not that that's bad, I'm just curious)? ok, i thought it was as you are saying, but i espected amanda to mailme with the results, but it did'd happen, and it didn't show the statistic either. So i changed my opinion and thought i should run it myself. John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
GNUtar problems on linux 2.2.16
Hello, I have a mini network with a host and a client (laptop) which is backed up with amanda for years. Now I upgraded the laptop to linux 2.2.16 (SuSE 7.0) and whith the result that I cannot backup the laptop anymore. Even worse, the entire amdump hangs and does not terminate anymore. What I have done so far is: Installed and compiled amanda 2.4.2 Installed, patched and compiled GNUtar-1.12 Using .amandahosts instead of BSD authorisation There are bad thing I can see: 1. Many warnings when compiling the server on function 'strncmp' Example: amadmin.c:1300: warning: integer overflow in expression 2. Warnings in amdump as: dumper: pid 2334 setsockopt(SO_RCVLOWAT): Protocol not available 3. In the clients sendsize.debug is an abort-message as --snip- sendsize: debug 1 pid 1851 ruid 37 euid 37 start time Fri Dec 29 13:28:36 2000 /usr/lib/amanda/sendsize: version 2.4.2 calculating for amname '/dev/hda5', dirname '/' sendsize: getting size via gnutar for /dev/hda5 level 0 sendsize: missing exclude list file "/usr/local/lib/amanda/exclude.gtar" discarded sendsize: running "/usr/lib/amanda/runtar --create --directory / --listed-incremental /var/lib/amanda/gnutar-lists/bop_dev_hda5_0.new --sparse --one-file-system --ignore-failed-read --totals --file /dev/null " sendsize: spawning "/usr/lib/amanda/runtar" in pipeline sendsize: argument list: "/bin/gtar" "--create" "--directory" "/" "--listed-incremental" "/var/lib/amanda/gnutar-lists/bop_dev_hda5_0.new" "--sparse" "--one-file-system" "--ignore-failed-read" "--totals" "--file" "/dev/null" "." /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./etc/mail/access.db: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./etc/mail/genericstable.db: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./etc/mail/mailertable.db: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./etc/mail/virtusertable.db: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./etc/postfix/virtual.db: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./root/.gnome/metadata.db: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./root/nsmail/.Entwürfe.summary: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./root/nsmail/.Gesendet.summary: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./root/nsmail/.Inbox.summary: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./root/nsmail/.Papierkorb.summary: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./root/nsmail/.Schablonen.summary: Bad file descriptor /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./root/nsmail/.Ungesendet.summary: Bad file descriptor Total bytes written: 60907520 /bin/gtar: Error exit delayed from previous errors . sendsize: pid 1851 finish time Fri Dec 29 13:28:36 2000 -snap-- Please tell me if you need more information. Any help is really appreciated Peter
Re: RV: amreport
On Fri, 29 Dec 2000, [iso-8859-1] Monserrat Seisdedos Nuñez wrote: hello every body: i'm testing amreport with an amanda run i have just done. i go to the /usr/local/etc/amanda/Diario where i have amanda.conf i type amreport segmentation fault(core dump) why?? Strange... are you sure the amreport binary you executed was built on the same machine or at least the same architecture as the machine you executed it on? -Mitch Yes, i got it from the same tar ball as the rest of the amanda. I'm not sure if I understand. Are you saying you built your amreport binary from the same source tree as the rest of the amanda binaries? Or are you saying you got a tar ball from someone that had all the binaries in it and you never built amanda from the source tree? -Mitch
Without index, why?
Hi folks, I built AMANDA without using --with-indexserver option since its default on local machine. On tape machine, the amanda.conf has the entry for indexdir, and "index yes" in the "define dumptype root-tar" "define dumptype user-tar" as well. Amanda should automaticlly produce index during backup. But none on our backup machine. What am I missing? I've got "501 No index records for host:backup. Invalid?" Error when I use amrecover. Thank you for you help. __ Do You Yahoo!? Yahoo! Photos - Share your holiday photos online! http://photos.yahoo.com/
Re: incremental backup without tape
I'd like to perform incremental backup of my servers leaving it on holding-disk Is it possible to do? How can I do it? Set tapedev to "/no/such/device". Alessandro Chiauzzi John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
RE: incremental backup without tape
If there is no tape in the drive, the backup will go to the holding disk (as long as it fits), and wait there for the tape and the `amflush` command. Note that this will not change the level of backup that AMANDA picks, and she may choose to do level 0 backups on some of your disks. the 'bump' option on the `amadmin` command may help, check the man page. -Original Message- From: alessandro chiauzzi [mailto:[EMAIL PROTECTED]] Sent: Friday, December 29, 2000 4:01 AM To: [EMAIL PROTECTED] Subject: incremental backup without tape I'd like to perform incremental backup of my servers leaving it on holding-disk Is it possible to do? How can I do it? Thanks a lot -- Alessandro Chiauzzi System Engineer Dialogos3 srl
Re: Without index, why?
On tape machine, the amanda.conf has the entry for indexdir ... What's in that directory? There should be one subdirectory for each client being backed up, then one subdirectory within each of those for each file system, then within each of those some datestamped gzip'd index files. If all that seems OK, what are the first 10 lines of a typical index file, e.g. "zcat xxx.gz | head"? Is one of the host level directories under index dir named "backup"? Or is it some alternate name for that host? John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: GNUtar problems on linux 2.2.16
What I have done so far is: Installed and compiled amanda 2.4.2 Installed, patched and compiled GNUtar-1.12 ... /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./etc/mail/access.db: Bad file descriptor Are you sure /bin/gtar is a patched version of 1.12? That error message is explicitly one that is corrected with the Amanda patch. There are bad thing I can see: 1. Many warnings when compiling the server on function 'strncmp' Example: amadmin.c:1300: warning: integer overflow in expression What compiler and options are you using? 2. Warnings in amdump as: dumper: pid 2334 setsockopt(SO_RCVLOWAT): Protocol not available I think you can ignore this. Peter John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: amdump
i amcheck my config file and i amdump it but with the tapedev="/dev/null" so it was copied to the holding disk ok. Don't do that. Use tapedev "/no/such/device". Some versions of Amanda will treat "/dev/null" as a debugging mechanism and throw the backup images away. but when i run amdump with the tapedev="/dev/nst0" the system hang, but the backup is done, i can see the files in the tape when i do tar: xvf /dev/nst0. ... If you can see things with "tar xvf /dev/nst0" then the backups were not done. Amanda tapes are not tar tapes. Tar should have complained. It is normal that the system hang when the amdump is runnig??? Obviously not. What do mean when you say the system hung? It went completely catatonic and you had to reboot? Or did just Amanda hang? Or something else? If the system hung, that's almost certainly a hardware problem. Start checking your cables, reseat the boards, etc. Did you amlabel the tape you tried to use? Was the tape brand new? What kind of tape drive are you using? John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: RV: amreport
i am using amanda 2.4.2 Are you using the real 2.4.2 release, or one of the beta's, or some RPM someone put together? Did you build from source? I include the output of the gdb. I don't understand anything. You didn't include the gdb output. ... but i espected amanda to mailme with the results, but it did'd happen ... That's because amreport is the one that generates the E-mail and it is failing. Once we get that fixed, you'll start getting the E-mail. I've discovered that one of the errors is generated because there isn't a tapelist file ... What error? Are you saying that if you create a tapelist file, amreport works? ... but i thought it was generated by amanda itself. Is that right??? Yes, that is right. It should have been created either by amlabel (have you run that?) or taper. i have the line: tapelist "/usr/local/etc/amanda/Diario/tapelist" , that is, in the same directory as amanda.conf Does the Amanda user have read/write/execute access to that directory? In other words, what is your Amanda user, what groups is it a member of, and what does "ls -lL /usr/local/etc/amanda/Diario" say? mo John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: GNUtar problems on linux 2.2.16
Hello John, Thank you for your really fast response. I am now some steps ahead, but not across the river yet... On 29-Dec-00 John R. Jackson wrote: What I have done so far is: Installed and compiled amanda 2.4.2 Installed, patched and compiled GNUtar-1.12 ... /bin/gtar: Read error at byte 0, reading 512 bytes, in file ./etc/mail/access.db: Bad file descriptor Are you sure /bin/gtar is a patched version of 1.12? That error message is explicitly one that is corrected with the Amanda patch. Good hint - /bin/tar was old, while /usr/local/bin/tar was new, but never used. Oh well There are bad thing I can see: 1. Many warnings when compiling the server on function 'strncmp' Example: amadmin.c:1300: warning: integer overflow in expression What compiler and options are you using? On the server i use gcc 2.7.2.3 with linux 2.2.13 configure options #!/bin/bash ./configure \ --prefix=/usr \ --bindir=/usr/bin \ --sbindir=/usr/bin \ --libexecdir=/usr/lib/amanda \ --sysconfdir=/etc \ --datadir=/usr/share \ --sharedstatedir=/usr/share \ --localstatedir=/var/lib/amanda \ --with-user=amanda \ --with-group=disk \ --with-rundump \ --with-config=bol \ --with-tape-server=bol \ --with-tape-device=/dev/nrmt0 \ --with-gnutar-listdir=/var/lib/amanda/gnutar-lists \ --without-bsd-security \ --with-buffered-dump \ --with-debugging=/tmp/amanda -- Currently I _CAN_ backup my slave, but deadly slow. The LAN show some activity every 10 seconds!!! On the server I can see processes like: 6133 pts/1S 0:00 | \_ sh /usr/bin/amdump bol 6139 pts/1S 0:00 | | \_ /usr/lib/amanda/driver bol 6140 pts/1S 0:00 | | \_ taper bol 6145 pts/1S 0:00 | | | \_ taper bol 6141 pts/1S 0:00 | | \_ dumper bol 6172 pts/1S 0:00 | | | \_ /usr/bin/gzip --best 6142 pts/1S 0:00 | | \_ dumper bol 6143 pts/1S 0:00 | | \_ dumper bol 6144 pts/1S 0:00 | | \_ dumper bol 6174 pts/1S 0:00 | \_ sh ./ldump 7505 pts/1S 0:00 | \_ sleep 10 How can I get rid of these 10 seconds? John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] Peter -- E-Mail: Peter Bollinger [EMAIL PROTECTED] Date: 29-Dec-00 Time: 18:45:37 This message was sent by XFMail --
Hardware problem?
Hi,
I dont know how to diagnose this problem. I first thought it
was that my tape was full but amanda tells me that its only using 78.3% of
the tape. I've tried different tapes and even changing the tapelist
file. This is really weird since I never had problem with amanda for more
than a year and a half. So I was wondering if someone can tell if this
is a hardware problem by this message or one of my assistants messed
around with the backup system =). Thanks in advance. Here is the message
I got:
Date: Fri, 29 Dec 2000 08:15:47 -0800 (PST)
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: EE AMANDA MAIL REPORT FOR December 29, 2000
These dumps were to tape Vol08.
Tonight's dumps should go onto 1 tape: Vol09.
FAILURE AND STRANGE DUMP SUMMARY:
taper: FATAL syncpipe_get: r: unexpected EOF
taper: FATAL syncpipe_get: r: unexpected EOF
taper: FATAL syncpipe_get: r: unexpected EOF
STATISTICS:
Total Full Daily
Dump Time (hrs:min)8:16 7:39 0:00 (0:10 start, 0:27
idle)
Output Size (meg) 15632.415632.40.0
Original Size (meg) 27445.927445.90.0
Avg Compressed Size (%)57.0 57.0--
Tape Used (%) 78.3 78.30.0
Filesystems Dumped 10 10 0
Avg Dump Rate (k/s) 583.5 583.5--
Avg Tp Write Rate (k/s) 581.7 581.7--
?
NOTES:
driver: WARNING: got empty schedule from planner
driver: WARNING: got empty schedule from planner
driver: WARNING: got empty schedule from planner
taper: tape Vol08 kb 16007872 fm 10 [OK]
?
DUMP SUMMARY:
DUMPER STATS TAPER
STATS
HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS
KB/s
-- --
--
coredump c1t0d0s0 03379113568 40.20:14 982.90:25
554.8
coredump c1t0d0s1 0 4928575 2887968 58.6 83:26 576.8 83:28
576.7
coredump c1t0d0s3 0 5514911 2732544 49.5 77:40 586.4 77:41
586.2
coredump c1t0d0s4 0 5939647 3542912 59.6 101:23 582.5 101:24
582.3
coredump c1t0d0s5 0 6075519 4317888 71.1 123:06 584.6 123:08
584.5
coredump c1t0d0s6 0 3175327 1470592 46.3 42:20 578.9 42:21
578.7
coredump c1t1d0s1 0 12985540384 31.13:28 193.71:13
557.1
eec0d0s0 0 826847 412064 49.89:29 724.3 11:51
579.6
eec0d0s3 0 1119039 490144 43.8 12:36 648.6 14:09
577.4
eec0d0s4 0 36111999488 27.53:28 477.32:58
559.3
(brought to you by Amanda version 2.4.0)
--
Isaac Saldana
Re: How to reuse a tape?
On Fri, 22 Dec 2000, David Wolfskill wrote: Date: Fri, 22 Dec 2000 13:48:31 -0400 (AST) From: The Hermit Hacker [EMAIL PROTECTED] I tried typing: %amadmin hub.org reuse Vol_1 amadmin: tape Vol_1 already reusable. yet amcheck hub.org tells me that I need a new tape ... Sounds as if the number of tapes amanda knows about that are already written to is less than the claimed tapecycle. Okay, when I originally set it up, I set it up with a 28 day cycle, figuring I'd pick up more tapes faster then I did :( If I set it down to, say, a 7 day cycle, and then decide to increase it as I acquire more tapes, will there be a problem, or is this not a problem to do?
Re: GNUtar problems on linux 2.2.16
Currently I _CAN_ backup my slave, but deadly slow. The LAN show some activity every 10 seconds!!! ... Sounds like some kind of networking problem. What happens if you try to ftp a reasonbly sized (10's of MBytes) file from the client to /dev/null on the server? Is there any indication the client backup is making progress, e.g. if you run lsof on gtar? What about on the server? Is the tape moving? Does "lsof -o" say taper is getting any data? Peter John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Hardware problem?
I dont know how to diagnose this problem. I first thought it was that my tape was full but amanda tells me that its only using 78.3% of the tape. ... All that means is that you wrote 78.3% of the "length" value in your tapetype. It does not mean you actually used that much of the physical tape (there is no easy way to know that). taper: FATAL syncpipe_get: r: unexpected EOF Most of these problems are fixed with the latest Amanda release (2.4.2). There might be more information at the end of the amdump.NN file. driver: WARNING: got empty schedule from planner This is very strange. I don't understand how driver could get an empty schedule yet the report indicates a bunch of stuff was done. Are your log.MMDD.NN files getting cycled, i.e. does there appear to be a new one for each run? Isaac Saldana John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: How to reuse a tape?
If I set it down to, say, a 7 day cycle, and then decide to increase it as I acquire more tapes, will there be a problem, or is this not a problem to do? I assume you are talking about your dumpcycle? This should be fine. It may take Amanda a couple of cycles to get the balance re-adjusted, but that's no big deal, especially since you'll be increasing the value. John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
GNU-Tar vs dump/restore ...
what are ppl more partial to using? I'm using dump right now, but there
are *alot* of files that I'm backing up there there is no reason to
backup, since restoring them is useless (ie. postgresql databases whose
state is questionable when backed up) ...
From reading the docs, I can't seem to find any way of doing an exclusion
list for dump, and the man page for dump re-enforces that ... so ...
... what are general opinions? ppl moving to gnu-tar vs dump, or just
living without the ability to do exclusions?
thanks ...
Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy
Systems Administrator @ hub.org
primary: [EMAIL PROTECTED] secondary: scrappy@{freebsd|postgresql}.org
Re: GNU-Tar vs dump/restore ...
what are ppl more partial to using? ... This comes up here every once in a while and is basically a religious issue. The points you bring up, as well as being able to do subsets of an entire file system and portability, are all good reasons to use GNU tar. I won't use it here because it alters the access time, and the other benefits don't outweigh that, but that's just my own situation. Marc G. Fournier John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: GNU-Tar vs dump/restore ...
On Fri, 29 Dec 2000, John R. Jackson wrote: what are ppl more partial to using? ... This comes up here every once in a while and is basically a religious issue. The points you bring up, as well as being able to do subsets of an entire file system and portability, are all good reasons to use GNU tar. I won't use it here because it alters the access time, and the other benefits don't outweigh that, but that's just my own situation. Oh, definitely a good point there ... another good one is that incrementals don't work either, do they, with GNU-Tar?
Re: GNU-Tar vs dump/restore ...
On Dec 29, 2000, The Hermit Hacker [EMAIL PROTECTED] wrote:
another good one is that incrementals don't work either, do they,
with GNU-Tar?
They do. That's the reason why Amanda requires GNU tar.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
Re: GNU-Tar vs dump/restore ...
On 30 Dec 2000, Alexandre Oliva wrote: On Dec 29, 2000, The Hermit Hacker [EMAIL PROTECTED] wrote: another good one is that incrementals don't work either, do they, with GNU-Tar? They do. That's the reason why Amanda requires GNU tar. okay, if that is the case, what exactly is required to switch over to gnu-tar? I don't need to do it for all my file systems, but have one in particular that I'd like to use it for ... and I'm assuming that I *can* intersperse as required? ie. use both dump and gnutar methods interchangeably?
Re: GNU-Tar vs dump/restore ...
On Dec 30, 2000, The Hermit Hacker [EMAIL PROTECTED] wrote:
okay, if that is the case, what exactly is required to switch over to
gnu-tar?
Select a dumptype that has GNUTAR as the backup program and force a
full backup of any disks whose backup program has changed.
I'm assuming that I *can* intersperse as required? ie. use both
dump and gnutar methods interchangeably?
Just make sure you don't attempt to create incrementals using a backup
program different from the one used for the last full. That's why you
should force a full backup whenever you switch.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
Re: GNU-Tar vs dump/restore ...
On Fri, Dec 29, 2000 at 08:26:58PM -0500, John R. Jackson wrote: what are ppl more partial to using? ... This comes up here every once in a while and is basically a religious issue. The points you bring up, as well as being able to do subsets of an entire file system and portability, are all good reasons to use GNU tar. I won't use it here because it alters the access time, and the other benefits don't outweigh that, but that's just my own situation. Using tar I only exclude ./tmp and the holding disk directory. Are there any problems backing up the FS containing the holding disk with dump (ufsdump in my case)? jon -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing [EMAIL PROTECTED] 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
