Following symlinks
Title: Message Hi, I have 2 problems on my hand with Amanda, and I was wondering if any of you would care to help a clueless newbie. First prob: I have a raid system running ReiserFS (which is why I am turning to Amanda, ditching ARCserve) where Mac users(via netatalk)would create tons of directories with funny characters like spaces and danish characters. I would like to back that up for some mysterius reason (duh!). Since the raid system is to large for 1 DLT tapeI can't dump the whole file system to tape. So I have to use gnutar and referencethe directories. You cannot put entries in the disklistfile with funny characters, soI tried to make a directory named .backup where I put symlinks to the directories called dump1, dump2 etc. This way I could manage the dumpsizes by gathering symlinks in the directories glancing at the directory sizes. The only thing is, now I have a bunch of tapes with symlinks on them and nothing else. I have also tried making hard links but that is not permitted, even though I am root and try to force it with the -F option. Second prob: It turns out after installing the Amanda from scratch with source compilation etc. that the system already had anAmanda RPM installed. Naturally, the first thing I did when I learned that was to do a 'rpm -e amanda'. But now I have pieces of the RPM installed and pieces of the Amanda source compialtion installed. How do I wipe out everything from both installations so I can have a clean install from source? I tried all kinds of make uninstall, make distclean, reinstalling and so on. I still have some amanda deamons wanting stuff one place and other daemons wanting the same stuff somewhere else... Any help would be appreciated... ;-Pete
Re: forcing a degraded backup
On Mon, May 06, 2002 at 04:18:08PM -0400, Jean-Louis Martineau wrote: On Mon, May 06, 2002 at 11:32:09AM +0200, Manuel Bouyer wrote: On Fri, May 03, 2002 at 03:23:06PM +0200, Manuel Bouyer wrote: Hi, I'm looking for a way to force a backup to degraded mode, to holding disk with 2.4.3b3. Reading the docs I didn't find anything obvious yet. Here's what I'm trying to do: I have a 14 slots DLT changer. With a runspercycle of 7, I have plenty of space left on tapes, so I'd like to put 2 days of backup on a tape. I'm trying to get amanda do a degraded backup to holding disk (incrementals only), which should be quite small (less than 5G), one day out of two. The next day this backup would be flushed to tape before the scheduled backups. Of course I'd like to have all this under the same config, so that peoples running amrecover don't have to know if it was a full or degraded backup. The problem is that I don't know how to make it do a degraded dump when there is a valid tape available. Any idea ? I found a way of doing it: setup a second configuration with a dummy tape device (I used /dev/null) with holding disk, infofile, logdir and indexdir pointing to the same place as the config with tape changer. This seems to work (I tried amrecover from both tape and holding disk). That's BAD, you can't use the same logdir for two configurations, logfile will get erased. You will get the same problem with indexdir if you have a disk in both config. I don't understand why ... note that both configs won't run at the same day (one day for backup to tape, next day a degraded backup). When I use 'history' in amrecover I get all the backups in the list, with the proper tapes. Also the log files are all there. Why not use one config and edit your amanda.conf everyday, that could be easily done by cron. I though at this, but I prefer static config files. This avoids problems in case the server goes down in the middle of a backup. Don't use /dev/null, your backup can go to it, use /dev/nosuchdevice instead. Ha, good point. A minor annoyance is that both amcheck and amdump complain about the invalid tape, amd amdump report claims that something went bad when it's what I wanted. Wish to amanda hackers: would be nice if there was a config option to make amanda work without tape at all, without complains :) Wish you will become an amanda hackers :) Well, I'm already a NetBSD developer. Days only have 24 hours :) -- Manuel Bouyer, LIP6, Universite Paris VI. [EMAIL PROTECTED] --
Question in case of disaster
Hi, I'm not in a disaster situation, fortunatelly, but I would like to test how to proceed, just in case. I run amanda-2.4.2p1-2. Supose an intruder break somehow the system and perform a rm -rf /. From this point, after a fresh reinstall of the system, how I can restore from tape, if no backup database is not availale anymore? I read Restoring Without AMANDA but some questions remains: (1) dd if=$TAPE bs=32k skip=1 of=/tape_content produce what in /tape_content? An ISO image that I can mount it with mount -o loop? (2) What is mentioned ufsrestore program? I tried to find what RPM (I ran RedHat) package belongs to but I did not find anything. My interest is to know ow to restore an entire directory and all it's subdirectories and all their files content in the shortest time possible after an unwanted disaster may occur. Thank you, Radu -- Radu Filip Network Administrator @ Technical University of Iasi [EMAIL PROTECTED] Information Technology and Communication Center http://socrate.tuiasi.ro/ [EMAIL PROTECTED] | http://ccti.tuiasi.ro/
RE: Backup NT share with amanda
you may want to test the SAMBA connection: smbclient nt-server\\share (The extra slashes are not a mistake. They are used to escape the backslashes under unix.) Just to save you guys some keystrokes, forward slash works equally well: smbclient //nt-server/share -- Ulrik Sandberg I got a similar problem with my nt-clients. Amcheck says, it could not connect, because the password was wrong. Smbclient connects fine with the same password. Didn't find out the reason. What I do as a workaround: Just before backing up, I mount the nt-shares on the backup-server using smbmount. So I can back them up with any constellation using GNUTAR. Maybe that helps you. Anyway I would be interested in a real answer. Maybe it has something to do with samba using encrypted/unencrypted passwords? Anybody got an idea?
Re: Question in case of disaster
My interest is to know ow to restore an entire directory and all it's subdirectories and all their files content in the shortest time possible after an unwanted disaster may occur. To do this in the shortest time possible the answer must be back up the amanda database files onto floppy disk on a regular basis, also test the floppy regularly. If your DB files won't fit onto a floppy consider CD or the tape your backing the other stuff on to. Although may have complicatyions with this method because the DB files will be open while amanda is running. In this scenario you only have to dig through the tape for the DB files restore them then restore as normal. This is the way we do it although touch wood we have not been faced with such a situation to date. Now that I've said that I know it's going to happen. David Flood Systems Administrator
Re: Seagate 4586 changer vs compression settings
Hi, one last resort would be using a big magnet on the tape and after that relabeling it. But be aware this erases every single bit ever written to the tape. one other thing i would try before is to look at the drive itself and check if there is a hardware switch to turn off compression. Many drives will read happily compressed tapes when compression is disabled, but they will use no compression anymore for writing. At least my STD224000N and the other drives from different vendors i have behave this way Christoph Gene Heskett wrote: Does anyone here have any idea how one would go about converting a DDS2 tape that been written by one of these with the DC on to a format that leaves the DC off? The problem is that once written to the tapes headers, it apparently cannot be switched back off. This appears to be in a header location on the tape that is not available to the outside world amlabel -f /config/ tapename slot # will dutifully read the old label if it exists, then rewind and write the new label just fine, but when it does the verify read after the write, the (*^ DC led comes right back on. I've even gone so far as to destroy the existing externally available header data with a dd from urandom, but this doesn't appear to touch the tape drives own, maintained on the tape, data. I'd like to have a little more control over the storage size of a tape than this allows, but it appears the only way to convert back is to toss these tapes and get fresh ones, and I have about 30 to replace. Ouch... :-(
Re: Question in case of disaster
On Tue, 7 May 2002 at 11:18am, Radu Filip wrote (1) dd if=$TAPE bs=32k skip=1 of=/tape_content produce what in /tape_content? An ISO image that I can mount it with mount -o loop? It is a file containing a backup image of a single disklist entry. What type of image depends upon what backup tool you are using. If you are using DUMP, then it's a dump image particular to the filesystem backed up (i.e. xfsdump for an XFS filesystem, dump for an ext2 filesystem on Linux, etc). If you use GNUtar, then it's a tarball. (2) What is mentioned ufsrestore program? I tried to find what RPM (I ran RedHat) package belongs to but I did not find anything. ufsdump/restore is the DUMP on Solaris for their UFS filesystem. You won't find it on Linux. My interest is to know ow to restore an entire directory and all it's subdirectories and all their files content in the shortest time possible after an unwanted disaster may occur. As someone else mentioned, the easiest way is to back up the amanda DBs (I tar 'em up each night and copy them to our RAID), which will then allow you to find out which tapes you need. Otherwise, you can read the header of each tape file via 'dd if=$TAPE bs=32k count=1 of=tapefile_header' at the beginning of each -- the header contains the filesystem info, date, backup level, and restore command. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: First run, how do I know it happened
On Mon, 6 May 2002 at 3:14pm, Andrew Falanga wrote Ok, something definitely happened. I can tell this by all the directories and files created under areas such as: /var/amanda/log and so forth. However, when I run amcheck config I get and error message that says, /var/amanda/log/log no such file exists (or something like that). Getting amcheck to pass should be the *first* thing you do, before amdump. Please post the exact amcheck error you are getting, and the relevant sections of amanda.conf. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: forcing a degraded backup
On Tue, May 07, 2002 at 10:18:44AM +0200, Manuel Bouyer wrote: On Mon, May 06, 2002 at 04:18:08PM -0400, Jean-Louis Martineau wrote: On Mon, May 06, 2002 at 11:32:09AM +0200, Manuel Bouyer wrote: On Fri, May 03, 2002 at 03:23:06PM +0200, Manuel Bouyer wrote: Hi, I'm looking for a way to force a backup to degraded mode, to holding disk with 2.4.3b3. Reading the docs I didn't find anything obvious yet. Here's what I'm trying to do: I have a 14 slots DLT changer. With a runspercycle of 7, I have plenty of space left on tapes, so I'd like to put 2 days of backup on a tape. I'm trying to get amanda do a degraded backup to holding disk (incrementals only), which should be quite small (less than 5G), one day out of two. The next day this backup would be flushed to tape before the scheduled backups. Of course I'd like to have all this under the same config, so that peoples running amrecover don't have to know if it was a full or degraded backup. The problem is that I don't know how to make it do a degraded dump when there is a valid tape available. Any idea ? I found a way of doing it: setup a second configuration with a dummy tape device (I used /dev/null) with holding disk, infofile, logdir and indexdir pointing to the same place as the config with tape changer. This seems to work (I tried amrecover from both tape and holding disk). That's BAD, you can't use the same logdir for two configurations, logfile will get erased. You will get the same problem with indexdir if you have a disk in both config. I don't understand why ... note that both configs won't run at the same day (one day for backup to tape, next day a degraded backup). When I use 'history' in amrecover I get all the backups in the list, with the proper tapes. Also the log files are all there. It works now because it's a new config, After tapecycle days, one config will erase the log and index of the other config. Why not use one config and edit your amanda.conf everyday, that could be easily done by cron. I though at this, but I prefer static config files. This avoids problems in case the server goes down in the middle of a backup. I think that running a small sed script to patch a static config file template before every amdump run is easier than using two configs. 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: My client backup is populating directory names, but not files
Yes, the Amanda user has permissions to read the raw devices (remember, the backup is working correctly on the server itself); Strangely, the dump sizes (and the size of the dump images) are consistent with the df output ... which would seem to indicate the data is there ... so why can't I see it -Mike Martinez -Original Message- From: Joshua Baker-LePain [mailto:[EMAIL PROTECTED]] Sent: Monday, May 06, 2002 1:45 PM To: Martinez, Michael - CSREES/ISTM Cc: '[EMAIL PROTECTED]' Subject: Re: My client backup is populating directory names, but not files On Mon, 6 May 2002 at 1:00pm, Martinez, Michael - CSREES/ISTM wrote On the other client, the backup procedure seems to go well - log files show okay. The problem comes in restoring. When using amrecover, it errors with connection reset by peer (as recorder on the server debug file). When using amrestore, an image is restored, and this image contains the first level (top level) directories from the client, but these directories are not populated with files! Any idea why files are not getting populated? I'm using a stock amanda install; with dump instead of tar, on RH 7.1 Are you sure the amanda user has permission to read the raw devices? Are the dump sizes reported by amanda consistent with the sizes reported by df? -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
unsubscribe
unsubscribe [EMAIL PROTECTED]
RE: My client backup is populating directory names, but not files
On Tue, 7 May 2002 at 7:59am, Martinez, Michael - CSREES/ISTM wrote Yes, the Amanda user has permissions to read the raw devices (remember, the backup is working correctly on the server itself); Whether or not the backup is working on the server is immaterial (in this case) to whether or not the backups from the client contain all the info they should. Strangely, the dump sizes (and the size of the dump images) are consistent with the df output ... which would seem to indicate the data is there ... so why can't I see it Does 'restore -tf $DUMP_IMAGE' list all the files? -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
RE: Backup NT share with amanda
Do you by chance have a /tmp/amanda/sendsize*debug file from the client? Looking where the failure comes from it may be when planner sends the sendsize request and the result was not what it was expecting to see. This may be one option to look at. Doug -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Monday, May 06, 2002 7:37 PM To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: RE: Backup NT share with amanda Two forms its successfully, but I cant backup the files. The error is the same. Thanks. From: Ulrik Sandberg [EMAIL PROTECTED] To: Doug Johnson [EMAIL PROTECTED] CC: [EMAIL PROTECTED], [EMAIL PROTECTED] Subject: RE: Backup NT share with amanda Date: Tue, 7 May 2002 00:56:48 +0200 (Västeuropa, sommartid) you may want to test the SAMBA connection: smbclient nt-server\\share (The extra slashes are not a mistake. They are used to escape the backslashes under unix.) Just to save you guys some keystrokes, forward slash works equally well: smbclient //nt-server/share -- Ulrik Sandberg _ Envíe y reciba su correo de Hotmail desde el móvil: http://mobile.msn.com
Re: Question in case of disaster
I have an rsync running to an alternate server so the index files are replicated to once the backup is completed. So I always have a fall back server. David Flood wrote: My interest is to know ow to restore an entire directory and all it's subdirectories and all their files content in the shortest time possible after an unwanted disaster may occur. To do this in the shortest time possible the answer must be back up the amanda database files onto floppy disk on a regular basis, also test the floppy regularly. If your DB files won't fit onto a floppy consider CD or the tape your backing the other stuff on to. Although may have complicatyions with this method because the DB files will be open while amanda is running. In this scenario you only have to dig through the tape for the DB files restore them then restore as normal. This is the way we do it although touch wood we have not been faced with such a situation to date. Now that I've said that I know it's going to happen. David Flood Systems Administrator -- _ Don Potter -- [EMAIL PROTECTED] Center for Human Genetics Duke University -_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_ Never knock on Death's door. Ring the doorbell and run (he hates that)
RE: My client backup is populating directory names, but not files
Yes, all the files are listed, both using restore -tf and amrecover.
Another thing I discovered is that the backup is not entirely good, on the
server as well as the client. I can restore some files, but not others, even
though these others may be listed in the index.
I'm getting checksum errors, as for example:
Load tape xxx now
continue? y
Checksum error 243736, inode 100 file directory file - name
unknown
./bin/df is not on the tape
That was when I was trying to restore bin/df as a test; and this /bin/df
is listed from amrecover.
So, I had several thoughts. Please confirm/disconfirm any you feel qualified
to do:
1. /sbin/dump is making mistakes. I'm currently testing Gnu Tar with a clean
slate.
2. My tape definition is incorrect for my drive? Maybe. I'm using a Sony
DDS-3 125 m (12 gig uncompresseD) tape, and I'm using hte following
definition which I took from the mailing list
define tapetype SDT-9000 {
comment Sony SDT-9000 DDS-3 DAT drive
length 12288 mbytes # 12GB native for 125 m tapes
filemark 0 kbytes
speed 1200 kbytes # kb/s sustained rate, compression disabled
}
3. Maybe the permissions on /var/lib/DailySet1/curinfo and index, are
incorrect? Don't know. I've got amanda.amanda drwxr-sr-x for both dirs
4. amcheck was complaining that subdirectories in those two above (like
_var, _home) didn't exist. But it said it would create them, so I figured
I'd let it
-Michael Martinez
-Original Message-
From: Joshua Baker-LePain [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, May 07, 2002 8:49 AM
To: Martinez, Michael - CSREES/ISTM
Cc: '[EMAIL PROTECTED]'
Subject: RE: My client backup is populating directory names, but not
files
On Tue, 7 May 2002 at 7:59am, Martinez, Michael - CSREES/ISTM wrote
Yes, the Amanda user has permissions to read the raw devices (remember,
the
backup is working correctly on the server itself);
Whether or not the backup is working on the server is immaterial (in this
case) to whether or not the backups from the client contain all the info
they should.
Strangely, the dump sizes (and the size of the dump images) are consistent
with the df output ... which would seem to indicate the data is there ...
so
why can't I see it
Does 'restore -tf $DUMP_IMAGE' list all the files?
--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University
Re: First run, how do I know it happened
[EMAIL PROTECTED] wrote: On Mon, 6 May 2002 at 3:14pm, Andrew Falanga wrote Ok, something definitely happened. I can tell this by all the directories and files created under areas such as: /var/amanda/log and so forth. However, when I run amcheck config I get and error message that says, /var/amanda/log/log no such file exists (or something like that). Getting amcheck to pass should be the *first* thing you do, before amdump. Please post the exact amcheck error you are getting, and the relevant sections of amanda.conf. Sorry, typo. I run amcheck with NO errors. I was running amreport as you'd suggested yesterday when I got that error. Andy
Re: First run, how do I know it happened
On Tue, 7 May 2002 at 8:38am, Andrew Falanga wrote Sorry, typo. I run amcheck with NO errors. I was running amreport as you'd suggested yesterday when I got that error. You need to specify the log file, e.g. 'amreport Daily -l /usr/local/adm/amanda/DailySet1/log.20020507.0 -f amreport.out' would re-generate this morning's report (for me) and stick it in the file amreport.out. See the man page for all the details. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
RE: My client backup is populating directory names, but not files
On Tue, 7 May 2002 at 10:17am, Martinez, Michael - CSREES/ISTM wrote
Yes, all the files are listed, both using restore -tf and amrecover.
Another thing I discovered is that the backup is not entirely good, on the
server as well as the client. I can restore some files, but not others, even
though these others may be listed in the index.
Uh oh.
I'm getting checksum errors, as for example:
*snip*
The *first* thing I'd do is make sure that your hardware is OK. Do a dump
by hand on the tape server to the tape drive, read it back, and confirm
that it's OK. If it doesn't work, then troubleshoot the hardware -- tape
drive, tapes, SCSI cables, SCSI chain, termination, color of goat. Are
you getting any errors in the logs?
Remember that amanda is, really, just a backup organizer. All backups
are actually done by system utilities (dump or tar) and the various device
drivers.
1. /sbin/dump is making mistakes. I'm currently testing Gnu Tar with a clean
slate.
Possibly, but my first suspicion is hardware -- probably the tape drive
since backups from both clients are failing (otherwise I'd also suspect a
hard drive).
2. My tape definition is incorrect for my drive? Maybe. I'm using a Sony
DDS-3 125 m (12 gig uncompresseD) tape, and I'm using hte following
definition which I took from the mailing list
define tapetype SDT-9000 {
comment Sony SDT-9000 DDS-3 DAT drive
length 12288 mbytes # 12GB native for 125 m tapes
filemark 0 kbytes
speed 1200 kbytes # kb/s sustained rate, compression disabled
}
No -- that's fine. About the only thing amanda really uses there is the
length parameter, in order to do estimates.
3. Maybe the permissions on /var/lib/DailySet1/curinfo and index, are
incorrect? Don't know. I've got amanda.amanda drwxr-sr-x for both dirs
I don't have the setgid bit. But this wouldn't cause the errors above.
4. amcheck was complaining that subdirectories in those two above (like
_var, _home) didn't exist. But it said it would create them, so I figured
I'd let it
Good choice.
Check your hardware.
--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University
Re: Following symlinks
Peter == Peter Normann [EMAIL PROTECTED] writes: Peter I have 2 problems on my hand with Amanda, and I was wondering if Peter any of you would care to help a clueless newbie. Peter First prob: Peter I have a raid system running ReiserFS (which is why I am turning Peter to Amanda, ditching ARCserve) where Mac users (via netatalk) would Peter create tons of directories with funny characters like spaces and Peter danish characters. I would like to back that up for some mysterius Peter reason (duh!). Peter Since the raid system is to large for 1 DLT tape I can't dump the Peter whole file system to tape. So I have to use gnutar and reference Peter the directories. You cannot put entries in the disklist file with Peter funny characters, so I tried to make a directory named .backup Peter where I put symlinks to the directories called dump1, dump2 Peter etc. This way I could manage the dumpsizes by gathering symlinks Peter in the directories glancing at the directory sizes. Peter The only thing is, now I have a bunch of tapes with symlinks on Peter them and nothing else. Peter I have also tried making hard links but that is not permitted, Peter even though I am root and try to force it with the -F option. reference the files as: /.backup/whatever/. The trailing /. will force the symlink to get dereferenced. ] ON HUMILITY: to err is human. To moo, bovine. | firewalls [ ] Michael Richardson, Sandelman Software Works, Ottawa, ON|net architect[ ] [EMAIL PROTECTED] http://www.sandelman.ottawa.on.ca/ |device driver[ ] panic(Just another NetBSD/notebook using, kernel hacking, security guy); [
RE: First run, how do I know it happened
On Tue, 7 May 2002 at 12:08pm, Morse, Richard E. wrote U... are you sure that it didn't send you the email? Try this: log on to your backup server, then su to the amanda user (you may need to be root to do this). Then type 'mail'. The report should be there. If it is, then you need to set up a .forward file, or create an alias in /etc/aliases. Actually, the easiest way to control who gets the amreport at the end of the run is via the mailto parameter in amanda.conf. Of course, you may still want to set up a .forward for the amanda account to get CRON errors, but you don't have to (you can do MAILTO in amanda's crontab as well). -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
RE: First run, how do I know it happened
U... are you sure that it didn't send you the email? Try this: log on to your backup server, then su to the amanda user (you may need to be root to do this). Then type 'mail'. The report should be there. If it is, then you need to set up a .forward file, or create an alias in /etc/aliases. If it isn't, then check to see if there is already a .forward file or an alias in /etc/aliases (and any other alias files specified in your sendmail.cf). If an alias is already set up, you need to check the mail as the user that it is aliased to. HTH, Ricky -Original Message- From: Joshua Baker-LePain [mailto:[EMAIL PROTECTED]] Sent: Tuesday 07 May 2002 11:00 AM To: Andrew Falanga Cc: amanda-users Subject: Re: First run, how do I know it happened On Tue, 7 May 2002 at 8:38am, Andrew Falanga wrote Sorry, typo. I run amcheck with NO errors. I was running amreport as you'd suggested yesterday when I got that error. You need to specify the log file, e.g. 'amreport Daily -l /usr/local/adm/amanda/DailySet1/log.20020507.0 -f amreport.out' would re-generate this morning's report (for me) and stick it in the file amreport.out. See the man page for all the details. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
File: driver problems
Hi
I'm still trying to get the file: driver stuff working, but I
think the docs aren't telling me everything I need to know...
I'm not exactly clear on how this is supposed to be set up.
I have a directory /tapes/, inside of which resides the
directories tapes01 to tapes20
My amanda.conf file has the following info:
runtapes 1
changerfile /usr/local/etc/amanda/foo/changer.conf
changerdev file:/tapes/
tapetype HARD-DRIVE
labelstr Foo-[0-9][0-9]
define tapetype HARD-DRIVE {
comment Hard Drive
length 102400 mbytes
}
And here's my changer.conf:
multieject 0
gravity 0
needeject 0
ejectdelay 0
statefile /usr/local/etc/amanda/foo/changer-status
firstslot 1
lastslot 20
slot 1 file:/tapes/tape01
slot 2 file:/tapes/tape02
slot 3 file:/tapes/tape03
slot 4 file:/tapes/tape04
slot 5 file:/tapes/tape05
slot 6 file:/tapes/tape06
slot 7 file:/tapes/tape07
slot 8 file:/tapes/tape08
slot 9 file:/tapes/tape08
slot 10 file:/tapes/tape10
slot 11 file:/tapes/tape11
slot 12 file:/tapes/tape12
slot 13 file:/tapes/tape13
slot 14 file:/tapes/tape14
slot 15 file:/tapes/tape15
slot 16 file:/tapes/tape16
slot 17 file:/tapes/tape17
slot 18 file:/tapes/tape18
slot 19 file:/tapes/tape19
slot 20 file:/tapes/tape20
When I do amcheck foo, I get the following error:
ERROR: /dev/null: rewinding tape: Inappropriate ioctl for device
(expecting a new tape)
Am I way off track here? Missing something, etc?
Do I need a tpchanger, and if so, what?
Thanks,
Ben S
--
RE: First run, how do I know it happened
Joshua Baker-LePain [mailto:[EMAIL PROTECTED]] wrote: Actually, the easiest way to control who gets the amreport at the end of the run is via the mailto parameter in amanda.conf. Of course, you may still want to set up a .forward for the amanda account to get CRON errors, but you don't have to (you can do MAILTO in amanda's crontab as well). Hmmm... but what if I want the report to go to many different people? Or people on different machines? I'm guessing that the amanda.conf method would work for this, I just wasn't sure, and for me, at least, I like separating the amanda issue of generating the reports from the office specific issue of who receives the reports... TMTOWTDI... Ricky
RE: First run, how do I know it happened
On Tue, 7 May 2002 at 12:19pm, Morse, Richard E. wrote Hmmm... but what if I want the report to go to many different people? Or people on different machines? I'm guessing that the amanda.conf method would work for From the sample amanda.conf: mailto $USER# space separated list of operators at your site So, e.g. mailto [EMAIL PROTECTED] [EMAIL PROTECTED] would work. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Seagate 4586 changer vs compression settings
On Tuesday 07 May 2002 07:08 am, Christoph Scheeder wrote: Hi, one last resort would be using a big magnet on the tape and after that relabeling it. But be aware this erases every single bit ever written to the tape. one other thing i would try before is to look at the drive itself and check if there is a hardware switch to turn off compression. Many drives will read happily compressed tapes when compression is disabled, but they will use no compression anymore for writing. At least my STD224000N and the other drives from different vendors i have behave this way Christoph This switch *was* turned on when I first time labeled this stack of tapes. It has since been turned off. But the tape drive itself, (true of all dats I think) maintains a copy of all this in a hidden from the user block of the tape, which explains why the drive must be able to read the tape before it will be placed in the 'ready' condition. If it finds the compression on when it reads this block, then it over-rides the dip switch settings. One doesn't see this until after the label has been written though. Its during the verify read that amlabel does to verify that it wrote the label correctly that the front panel DC led comes back on, and it stays on for all subsequent writes to the tape. In other words, once turned on, it cannot be turned off later. I've even let it accept the tape, then wrote a compression off to the drive which does turn the drive DC led off, but then amdump, in its infinite wisdom, wants to make sure its the right tape, and it comes back on as the label is read by amdump. Degausing the tape: I tried that on 2 of these, which are now in the wastebasket, amlabel cannot even read them, IO error exits being the prefered method of handling that, likewise dd also exits with an io error. I assume thats because this MRS block of data on the tape has been wiped as the drive will spend maybe 3 minutes repeatedly rewinding and retrying to read the tape before tossing the error out. An 'mt -f /dev/nst0 erase also exits with an io error on such an degauser erased tape. One thing I haven't tried, but will, is to write the DC off, then eject the tape immediately, which will rewrite that hidden block before its ejected. That *might* do it. Food for thought anyway. [snip] -- Cheers, Gene (from the salt mine)
Re: File: driver problems
On Tue, 7 May 2002 at 12:16pm, Ben Snyder wrote runtapes 1 changerfile /usr/local/etc/amanda/foo/changer.conf changerdev file:/tapes/ tapedev contains the file: line, not changerdev, at least according to amanda(8). -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
RE: My client backup is populating directory names, but not files
Okay I've narrowed down the problem. Orginally I was having problems restoring data (checksum errors; missing files). I was using dump, so I switched to gnutar. Gnutar worked better, in that I was able to restore files. But when I went and did a diff or a cmp, many of the restored files do not match their originals: random byte-replacement; eg. the letter g intead of e; or the character of - instead of /; and so on. I stuck in a new cleaning tape; ran it through; stuck in a brand new tape; and did a direct tar to tape of one of my filesystems. I restored it back - same problem; lack of integrity in the data. Random bytes not pulling over -mike
RE: My client backup is populating directory names, but not files
In reference to this thread, in which it has been determined that writing to/from my tape drive is producing some loss of integrity, (mis-translation of random bytes, occurs both with tar, and dump, and using amanda) does anyone here know how to troubleshoot a SCSI tape drive or whether it is in fact necessary to troubleshoot it in this case, or whether I just chalk it up to a bad drive and get a new one. The drive is a HP SureStore DAT24.
