Re: [off-topip] Better Backup Media
On Tue, Jun 21, 2005 at 09:20:45AM -0400, Joshua Baker-LePain wrote: As long as we're talking drives and media (and I'm in the market), does anyone have any experience with S-AIT? According to my vendor, it has the same smart compresser as AIT and a capacity of 500GB/tape (vs. 400GB for LTO-3). LTO-3 looks like it may be the winner on price, but if there's a good reason to go S-AIT, I'm not averse to doing so. I don't have any SAIT experience, but the last time I looked, the big limitation for me was that nobody but Sony was selling a library with SAIT drives in it, and Sony was only selling them in their high end tape silos. That may have changed since the last time I looked. I seem to remember maybe seeing a Spectra Logic box with a SAIT drive in it at a conference sometime recently. I know that Overland isn't offering them in their Neo libraries, which is where I'd really like to see one. If I replace my Overland it'll probably be with another. It's been highly reliable. Right now SAIT looks to me like it'll probably win out over LTO for density and speed in the near term and in the long term, but helical scan has always been harder to sell than linear tape. The AIT technology has been highly reliable for me and seems to be as reliable for everyone else as well. AIT may finally be the technology that gets helical scan into the real enterprise market, but until you see an AIT derivative in an STK silo for instance, it seems like SAIT is going to remain a specialty product. Who knows, maybe this has all changed in the past few months, or is in the process of changing now. AIT-4 sure looks pretty though. Brandon -- Pseudo-Random Googlism: birthday is here; commemoration to continue
Re: [off-topip] Better Backup Media
On Mon, Jun 20, 2005 at 08:17:55PM -0300, Fabricio Luiz Machado wrote: Hi. Some of my DDS-4 tapes are bad (I/O error). So... i?m thinking if these tapes are trustworthy... :-/ Are there another better solution, that I can trust more than DDS-4 ? Maybe DLT ? DLT has a good track record, but is expensive and been obsoleted by newer technologies. A lot of us here speak very highly of AIT, and it's positioned to replace DDS in the market. I'd look into AIT. Brandon -- Pseudo-Random Googlism: sunshine is on a rage in japan
--atime-preserve: help is on the way
If you're like me, you have a heckuva lot more data to back up than a single tape can contain. You also probably have enough different architectures around that multiple operating system-specific dump formats are unattractive to you, even if you could dump entire filesystems and fit them on a tape. So what do you do? You use GNU tar, like everybody else. And why does this disappoint you? Because it's almost 2005 and you're still clobbering atime with every nightly backup. How lame. How terribly, terribly lame. Three and a half decades of unix and we still haven't got that fixed? Not anymore. Maybe. Eventually. At least in open source unixes. The Linux 2.6.8 kernel introduced an O_NOATIME option to the open(2) syscall which, when requested, will mean that subsequent read(2)s to the fd will not alter the atime of the file. O_NOATIME is only granted to the user who owns the file and the superuser. What remains to make GNU tar backups that don't clobber atime a reality? * A GNU tar patchset that teaches --atime-preserve about O_NOATIME and uses it opportunistically over the utime(2) method. * Probably a backport of O_NOATIME to Linux 2.4 for the [ likely millions of ] systems not ready to move to 2.6. * Kernel patchsets for your favorite BSD. This annoys me enough that I'm willing to work on the above list. However, my project list is such right now that it is likely several months before I will be far enough down the list to concentrate effort on improving our backup infrastructure and can justify spending work hours hacking at this. I happen to know there are tons of amanda users on this list whose sole business and livelihood is backups. This message is intended to let you know that this is quite possible to achieve today, and if backups are your business, it might be worth something to you to spend your time improving these tools. In other words, if you care about this as much as I do and can work on this sooner than I can, the work is out there waiting to be done and I'd love to see someone jump on it. I also know there are amanda developers out there who already have relationships with projects like GNU tar and might be able to exert some influence to see these features committed and released. Backups don't have to suck. Brandon -- Pseudo-Random Googlism: winter is typically known for hibernation periods
Re: Amanda users at LISA next week?
On Thu, Nov 11, 2004 at 09:58:49AM -0600, Frank Smith wrote: Anyone else on the list planning on attending LISA in Atlanta next week? It would be nice to be able to meet some of the people who's names I see on the list, and possibly have an unofficial BoF on Amanda to swap ideas and war stories. I will be at the conference, though fairly busy so unsure of how much time I will have for an Amanda pow-wow. FWIW, Aeleen Frisch is reportedly covering Amanda in her half-day Monday afternoon tutorial Beyond Shell Scripts: 21st-Century Automation Tools and Techniques[0]. If you want to coordinate some sort of meetup I will try to make it. [0] - http://www.usenix.org/events/lisa04/training/tutmon.html#m12 Brandon -- Pseudo-Random Googlism: election is ground zero in cultural wars
Re: amrecover failing
On Mon, Oct 18, 2004 at 01:26:24PM -0400, [EMAIL PROTECTED] wrote: I'm trying to use amrecover to restore a users homedir, and keep getting the following error when I answer 'Y' to the 'Load tape now' question: EOF, check amidxtaped.timestamp.debug file on space-monster.permabit.com. amrecover: short block 0 bytes UNKNOWN file amrecover: Can't read file header extract_list - child returned non-zero status: 1 Whenever I have seen this error from amrecover it is because the tape you are restoring from has not been rewound prior to running amrecover. Load the correct tape and then use 'mt -f /dev/tapedevice rewind' to make sure the tape is rewound before letting amrecover run and see if that helps you. HTH, Brandon -- Pseudo-Random Googlism: fall is a good time to focus on herd health
Re: fake install path
On Mon, Mar 01, 2004 at 10:10:13PM +0100, Paul Bijnens wrote: Jonathan Dill wrote: I vaguely recall that there is variable that you can pass to 'make' to install to a different root, similar to what happens during building a binary RPM, for example: make install VARIABLE=/tmp/amanda-2.4.4p2 The prefix= can be specified to override the configure prefix like: make install prefix=/tmp/amanda-2.4.4p2 This is not what you want, Jonathan. DESTDIR is the correct variable to overload for package building at the make install stage. PREFIX is defined at configure time as the install location. DESTDIR will be prepended to PREFIX at make install time. HTH, Brandon D. Valentine -- [EMAIL PROTECTED] http://www.geekpunk.net Pseudo-Random Googlism: february is national pet dental health month
Re: Changing Amanda server
On Wed, Feb 11, 2004 at 09:22:14PM -0500, Robert L Becker Jr wrote: I haven't seen this question in the list or archives. How do I gracefully move amanda databases from one (Irix) server to a new one (Linux). I expect to handle the build and configuration on the new host. I'll be moving our Exabyte jukebox from old to new host. Once the hardware and build dances are done, can I just (translate and?) move some files from the old server to the new one, keeping the same catalogs etc to access our current inventory of backups? If I understand what you're saying right, yes. The amanda database is just plain text files. Move the directories $PREFIX/etc/amanda/$CONFIG/(curinfo|index|log) to the corresponding location on the new server. Brandon D. Valentine -- [EMAIL PROTECTED] http://www.geekpunk.net Pseudo-Random Googlism: february is heart health month in canada
Re: List issues
On Mon, Jan 19, 2004 at 07:44:49PM -0700, Justin Gombos wrote: How are you folks handling it? I sort amanda-users based on the Sender header. Most older Majordomo installations don't offer any of the List-* or X-* headers, but Sender has been available via Majordomo for some time. Here's procmail: :0: * ^Sender: owner-amanda-users amanda-users HTH, Brandon D. Valentine -- [EMAIL PROTECTED] http://www.geekpunk.net Pseudo-Random Googlism: love is the devil caged on canvas
Re: Debian/Woody backport of 2.4.4p1 ?
On Mon, Dec 15, 2003 at 02:59:44PM -0500, Kurt Yoder wrote: Brandon D. Valentine said: % perl -p -i -e 's/debhelper \(= 4.1.16\)/debhelper/' debian/control (this step is necesary because the Debian amanda package maintainer uses the mess that is debhelper and it has marked itself as dependent on a version of debhelper that isn't in woody, but it works just fine with the version that is in woody) When I did this on my stable Amanda server a few weeks ago, I had trouble with the version of debhelper. The dpkg-buildpackage depended on debhelper from testing. So I had to disregard dependency problems using fakeroot dpkg-buildpackage -D -us -b -d (extra -d). Kurt, One of my steps solved the dependency problem. ;-) Brandon D. Valentine -- [EMAIL PROTECTED] http://www.geekpunk.net Pseudo-Random Googlism: tejas is the kinetic force
Re: Debian/Woody backport of 2.4.4p1 ?
On Fri, Dec 12, 2003 at 04:06:39PM -0500, [EMAIL PROTECTED] wrote: Does anyone know if there's a backport of amanda 2.4.4p1 to Debian Woody? I could use the packages in testing, but the require an upgrade of libc, which I really want to avoid on my production servers right now. Assuming your user is in the src group, testing is in your sources.list, you've run apt-get update recently and you've got fakeroot installed: % su - # apt-get build-dep amanda # exit % cd /usr/src % apt-get source amanda % cd amanda-2.4.4p1 % perl -p -i -e 's/debhelper \(= 4.1.16\)/debhelper/' debian/control (this step is necesary because the Debian amanda package maintainer uses the mess that is debhelper and it has marked itself as dependent on a version of debhelper that isn't in woody, but it works just fine with the version that is in woody) % dpkg-buildpackage -b -uc -rfakeroot % su - # dpkg -i /usr/src/amanda-*.deb And now Amanda 2.4.4p1 is installed, built against Woody. Use the force, Luke. HTH, Brandon D. Valentine -- [EMAIL PROTECTED] http://www.geekpunk.net Pseudo-Random Googlism: programming is the new rock 'n' roll
Re: [FYI] Heterogenous system restore
On Tue, Oct 07, 2003 at 05:40:49PM +0200, Nicolas Ecarnot wrote: The restore program of freebsd can't read the dumps made by my linux client. Google archives seems to say that this comes the filesystems : ext2 on linux vs ufs on freebsd. This is correct. FWIW, gnutar dumps don't suffer this problem. Unless atimes are somehow important to your application, I recommend that just about everyone use gnutar these days, especially in a heterogenous environment where you can't always count on the platform you made the dump on being available at restore time. The solution : [snip] Any comments welcome. You are aware you can run amrecover remotely, aren't you? HTH, Brandon D. Valentine -- [EMAIL PROTECTED] http://www.geekpunk.net Pseudo-Random Googlism: tejas is an excellent choice for happy hour
Re: planner problem
On Fri, Aug 15, 2003 at 02:53:00PM -0400, Shashi Kanbur wrote: Im running AMANDA 2.4.4p1 on Debian Linux systems. amcheck returns no problems. However, on some machines, when I run amdump, I get in the amdump log planner: time 1800.276: error result for host nova disk /data: Estimate timeout from nova planner: time 1800.276: error result for host nova disk /home: Estimate timeout from nova planner: time 1801.626: error result for host messier disk /data: Estimate timeout from messier even though these machines and disks are in the disklist file and amcheck returns no errors. Other machines work fine. Any ideas on how to solve this? Bump up the etimeout value in your amanda.conf. Read the amanda(8) manpage for more information. Brandon D. Valentine -- [EMAIL PROTECTED] http://www.geekpunk.net Pseudo-Random Googlism: unix is the answer
Re: Weekly amflush
On Wed, Jul 09, 2003 at 09:26:05AM -0600, Steven J. Backus wrote: I'd like to do my daily dumps to the holding disk, then amflush it to the tapes over the weekend. The suggestion I've heard here is to leave out the tapes, then put them in on Friday night, but what if I want to take Friday off? Is there a way to do this automatically? It might be more effort than you want to invest but you could write a fake changer script which would lie to Amanda six days a week and then every Friday tell Amanda it's found the correct tape. Then you can change the tape any time during the week you want as long as its before Friday. Brandon D. Valentine -- [EMAIL PROTECTED] http://www.geekpunk.net Pseudo-Random Googlism: peace is a mental construct
Re: this is one I haven't seen
On Fri, Jun 20, 2003 at 05:23:15PM -0500, FM Taylor wrote: After an upgrade all of my /dev/rmt/devices are now /dev/scsi/sequential/devices. That in itself was no problem. However, I am now getting this strange error, and I don't know how to fix it. tape_rdlabel: tape open: 0: No such file or directory What did I miss? According to your mail headers you sent this from a Windows98 machine, but I doubt that that is the platform you are running amanda on. How about letting us in on that secret? You also might consider telling us what command you are getting that tape_rdlabel error from. How about telling us what you set your tapedev parameter in amanda.conf to? Brandon D. Valentine -- Systems Administrator Center for Structural Biology Vanderbilt University
Re: problems using amanda with xinetd
On Tue, Jun 10, 2003 at 04:36:23AM -0400, Gene Heskett wrote: xinetd isn't a pos, its quite a bit more secure, and less wastefull of system resources than inetd because things don't get started at boot time and left around in case they are needed, they are started on demand, and killed when the demand is gone. Its also had a couple of security related updates fairly recently and the version I have installed is now 2.3.11 IIRC. If yours is older, I'd get the latest before I re-installed it. Actually, Gene, I'd have to take issue with this. xinetd does not leave any fewer processes hanging around than inetd does. In fact, they are functionally equivalent pieces of software. The primary difference between the two is that xinetd wastes inodes on your filesystem. =) There has to be some reason the services won't start, so please post an 'ls -l' of the /usr/local/libexec directory. Also an 'ls -l' of the amanda src directory, and a 'cat' of your configuration script. Since amanda works properly for Mike under inetd and not under xinetd I would suspect a problem with his xinetd build. My question to Mike is, if it works under inetd, why are you bothering with the PITA that is xinetd? ;-) Brandon D. Valentine -- Systems Administrator Center for Structural Biology Vanderbilt University
Re: Backing up MySQL tables
On Wed, Apr 02, 2003 at 12:27:34PM -0600, Alex Thurlow wrote: I need to start backing up MySQL tables on one of my machines, but I've heard that to you run the risk of table corruption if you just copy the mysql directory while MySQL is running. Stopping the database is not an option, and the only solution I've come up with is having another script do a mysql dump and then backing up that. It seems like a waste of time and resources to basically do 2 backups though. I was just wondering what anyone here does for MySQL backups on their own systems? Cron up mysqldump to happen just before the amanda run. Brandon D. Valentine -- Systems Administrator Center for Structural Biology Vanderbilt University
Re: Restoring backups
On Wed, Apr 02, 2003 at 03:29:38PM -0500, Keith wrote: We have Amanda configured to back up to disk files and have that working fine (maybe) and with samba/windows directories and gnutar. Anyway it appears to be working. But I can't get a restore to work. My test file is: /usr/backup/amandadumps/tape04/data/2.JESSIE.__Html_za__setup$.1 and its 65536 bytes. From the directory I'm trying: amrestore 2.JESSIE* It says: amrestore: 0: restoring JESSIE.__Html_za__setup$.20030319.1 And it creates a file: JESSIE.__Html_za__setup$.20030319.1 but with 0 bytes. What am I doing wrong or not doing right? You can't restore a level 1 without first restoring its level 0. Also, you need to read amrestore(8). To quote: Unless -p is used, candidate backup images are extracted to files in the current directory named: hostname.diskname.datestamp.dumplevel You will want to use the -p option to amrestore to pipe its output into tar in order to get the files back. Brandon D. Valentine -- Systems Administrator Center for Structural Biology Vanderbilt University
Re: Desired options.
On Tue, Apr 01, 2003 at 11:03:08AM -0500, Keith Higgs wrote: That said, there are lots of opportunities for additional configuration options. Our single tape drive system is physically in a secure area that is staffed by our central IT department. I have full access to the room but, we need to work out a process/procedure for their over-night operator to change out the tapes for us on a regular schedule. I would really like options to send tape administration reminders to a specified email address as part of the amcheck process. I could then schedule amcheck to run long enough ahead the call to amdump for the operator to change the tape if needed. Not to set aside your suggestion, because I think it's a worthwhile one, but here is a potential workaround which you could implement to solve your problem in the near term: 1) Set mailto in amanda.conf to an address where a procmail script is listening out of aliases(5). 2) In that procmail script do some initial error checking along the lines of does this look like an amanda report? did it come from the amanda user? and redirect anything that doesn't look like a legitimate amanda report to /dev/null or to the postmaster with an informative error message. 3) Test to see if a line in the message body matches some regex devised to only pick out tape errors. 4) Pipe a copy of any messages which match the rule into a script which automatically parses them to find out which tape amanda is looking for and sends a form mail to the overnight operator asking him to insert that tape. 5) Send the report on to the real amanda administrator (you). Or, alternately, talk someone into buying you a library. ;-) Cheers, Brandon D. Valentine -- Systems Administrator Center for Structural Biology Vanderbilt University
Re: Frontend , UI for amanda ?
On Mon, Mar 31, 2003 at 07:23:29PM -0500, Bort, Paul wrote: On Mon, Mar 31, 2003 at 05:00:49PM -0500, Thomas Hu wrote: Gene made some points but not all I agree with. Friendly user interface (not necessary fancy GUI) is a measure of all good quality software products. Amanda is not beyond the scope of this view. As always, it's open source. If the lack of a GUI bothers you enough, you can write one, or hire someone to write it for you. The software scratches the collective itch. Thomas, Bernd, et al: 1) This discussion probably belongs on amanda-hackers. 2) As I see it, amanda is in no way obligated to be all things to all people. In fact, that is a rather stupid way to approach software design. Software should be primarily concerned with solving one specific problem and doing the best job of that as is reasonably possible. See UNIX for details. 3) You seem to have a narrow view of the concept of user friendliness. An easy to use interface which leads the user to take the wrong action is far less user friendly than an interface with a learning curve which empowers the user with the knowledge that he is taking the correct action. It is my opinion that amanda's current interface is sufficiently user friendly to enable me as the administrator to do my job and know that it is being done in a manner which is verifiably correct. The data I protect is worth many hundreds of thousands of dollars in laboratory materials, instrument time and manhours. It is far more valuable for me to know that the data was backed up correctly than to be able to use a mouse to do so. Where data integrity is concerned my job is on the line as are the academic careers of many of my colleagues. I take that very seriously. Were I doing this job in a corporate environment I would take it equally seriously. 4) Code talks. Amanda is open source. So far nobody who has complained about the lack of a GUI has ponied up to write one. Will you be the exception? Experience makes me doubt that you will. If that offends you, prove me wrong. 5) As far as I am concerned there is _ONE_ place in the entire amanda model where a GUI would be useful. The GUI is not useful for me as the administrator. A GUI for amanda could be quite useful however for enabling individual users to schedule their own restores. However, in order for such an application to be possible amanda would need to overhauled to support some concept of user accounts and permissions so that individual users could be granted permission to restore their own data. That's a lot of work and should probably happen concurrently with a transition away from rhosts style auth(orization|entication) so that amanda transitions to a uniform model for such. Brandon D. Valentine -- Systems Administrator Center for Structural Biology Vanderbilt University
Re: tape drive opinions
On Sat, Mar 29, 2003 at 10:58:04PM -0500, Jeremy L. Mordkoff wrote: I need a 80 GB+ tape drive for a Dell Poweredge 2500 running Redhat 8.0 and Amanda This question seems to come up fairly regularly. I will direct you to the answer I have given here before: http://marc.theaimsgroup.com/?l=amanda-usersm=103133231011414w=2 HTH, Brandon D. Valentine -- Systems Administrator Center for Structural Biology Vanderbilt University
Re: planner bug?
On Sat, Mar 29, 2003 at 10:59:46PM -0500, Jeremy L. Mordkoff wrote: I have 20 GB (uncompressed) tape drive. I have a 23 GB partition. My compression ratio averages well over 50% so a full will fit on the tape, but Amanda refuses to even try I am unable to read your mind through your tinfoil hat. You will have to provide me with details. I am naturally inclined to suspect operator error over an actual bug in the planner. If you want someone to investigate the problem I suspect you submit your amreport output, the contents of your amanda logfile for the session in question, and the contents of the amandad.debug file from the client. Perhaps you should read this: http://www.catb.org/~esr/faqs/smart-questions.html Brandon D. Valentine -- Systems Administrator Center for Structural Biology Vanderbilt University
Re: Make amanda unload tape drive?
On Thu, 9 Jan 2003, John Oliver wrote: Can amanda send a signal to my DLT IV to unload the tape after backups are finished, so whoever goes back to swap tapes can just do it without waiting to manually unload? An excerpt from the Amanda user's crontab on my amanda server: 45 0 * * 2-6 /usr/local/sbin/amdump Daily /usr/bin/eject -q /dev/nst0 /dev/null That's all it takes. The '' will ensure that the eject command only executes if the amdump run returns zero (meaning it completed successfully). The redirection of stdout/stderr into /dev/null ensures you don't get obnoxious emails from cron following a run where there was no tape in the drive and so nothing to eject. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: make w/ amanda ignoring DESTDIR
On Tue, 17 Sep 2002, Jean-Louis Martineau wrote: It works on IRIX with GNU make It doesn't works on IRIX with /sbin/make /sbin/make will works if you remove all the DESTDIR = line from the Makefile. Thanks, Jean. Your suggestion made me do some further digging. IRIX's make is among a handful of variants which do not allow command-line variable definitions to override those that preexist in the Makefiles. It seems like the solution to this that the GNU people took was to remove all instances of DESTDIR = from their Makefile.in's. Maybe I'll submit some patches to amanda-hackers to fix it. It's a niggling little bug but if a one-line patch to each of the Makefile.in's will fix it then it's worth people not running into the problem in the future. An explanation of the problem is here: http://sources.redhat.com/ml/bug-automake/1999/msg00267.html -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: make w/ amanda ignoring DESTDIR
On Tue, 17 Sep 2002, Jean-Louis Martineau wrote: I would appreciate if many people try it, if it break nothing, I will use automake-1.5 to build the stable 2.4.3 Thanks, Jean. I'll test the builds here on my IRIX and Linux machines and report back. I'll just hand edit the Makefiles to remove 'DESTDIR =' so I can get on with building my 2.4.2p2 packages since I'm waiting on the 2.4.3 stable release to move over to it in any more than a developmental environment. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: make w/ amanda ignoring DESTDIR
On Tue, 17 Sep 2002, Jean-Louis Martineau wrote: The patch is not useful, the Makefile.in are automaticaly generated from the Makefile.am by automake. I made one anyway in case I ever need to build 2.4.2p2 packages again on IRIX. It's only 8k so I've attached it for the archives in case anyone else finds it useful. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED] diff -ru amanda-2.4.2p2.orig/Makefile.in amanda-2.4.2p2/Makefile.in --- amanda-2.4.2p2.orig/Makefile.in Tue Apr 3 14:55:47 2001 +++ amanda-2.4.2p2/Makefile.in Tue Sep 17 17:58:18 2002 @@ -32,8 +32,6 @@ includedir = @includedir@ oldincludedir = /usr/include -DESTDIR = - pkgdatadir = $(datadir)/@PACKAGE@ pkglibdir = $(libdir)/@PACKAGE@ pkgincludedir = $(includedir)/@PACKAGE@ Only in amanda-2.4.2p2/amplot: Makefile diff -ru amanda-2.4.2p2.orig/amplot/Makefile.in amanda-2.4.2p2/amplot/Makefile.in --- amanda-2.4.2p2.orig/amplot/Makefile.in Tue Apr 3 14:55:43 2001 +++ amanda-2.4.2p2/amplot/Makefile.in Tue Sep 17 18:01:55 2002 @@ -34,8 +34,6 @@ includedir = @includedir@ oldincludedir = /usr/include -DESTDIR = - pkgdatadir = $(datadir)/@PACKAGE@ pkglibdir = $(libdir)/@PACKAGE@ pkgincludedir = $(includedir)/@PACKAGE@ Only in amanda-2.4.2p2/amplot: amcat.awk Only in amanda-2.4.2p2/amplot: amplot.sh Only in amanda-2.4.2p2/changer-src: Makefile diff -ru amanda-2.4.2p2.orig/changer-src/Makefile.in amanda-2.4.2p2/changer-src/Makefile.in --- amanda-2.4.2p2.orig/changer-src/Makefile.in Tue Apr 3 14:55:43 2001 +++ amanda-2.4.2p2/changer-src/Makefile.in Tue Sep 17 18:02:06 2002 @@ -34,8 +34,6 @@ includedir = @includedir@ oldincludedir = /usr/include -DESTDIR = - pkgdatadir = $(datadir)/@PACKAGE@ pkglibdir = $(libdir)/@PACKAGE@ pkgincludedir = $(includedir)/@PACKAGE@ Only in amanda-2.4.2p2/changer-src: chg-chio.pl Only in amanda-2.4.2p2/changer-src: chg-chs.sh Only in amanda-2.4.2p2/changer-src: chg-manual.sh Only in amanda-2.4.2p2/changer-src: chg-mtx.sh Only in amanda-2.4.2p2/changer-src: chg-multi.sh Only in amanda-2.4.2p2/changer-src: chg-rth.pl Only in amanda-2.4.2p2/changer-src: chg-zd-mtx.sh diff -ru amanda-2.4.2p2.orig/client-src/Makefile.in amanda-2.4.2p2/client-src/Makefile.in --- amanda-2.4.2p2.orig/client-src/Makefile.in Tue Apr 3 14:55:44 2001 +++ amanda-2.4.2p2/client-src/Makefile.in Tue Sep 17 17:58:52 2002 @@ -34,8 +34,6 @@ includedir = @includedir@ oldincludedir = /usr/include -DESTDIR = - pkgdatadir = $(datadir)/@PACKAGE@ pkglibdir = $(libdir)/@PACKAGE@ pkgincludedir = $(includedir)/@PACKAGE@ Only in amanda-2.4.2p2/client-src: patch-system diff -ru amanda-2.4.2p2.orig/common-src/Makefile.in amanda-2.4.2p2/common-src/Makefile.in --- amanda-2.4.2p2.orig/common-src/Makefile.in Tue Apr 3 14:55:44 2001 +++ amanda-2.4.2p2/common-src/Makefile.in Tue Sep 17 17:59:03 2002 @@ -34,8 +34,6 @@ includedir = @includedir@ oldincludedir = /usr/include -DESTDIR = - pkgdatadir = $(datadir)/@PACKAGE@ pkglibdir = $(libdir)/@PACKAGE@ pkgincludedir = $(includedir)/@PACKAGE@ Only in amanda-2.4.2p2/common-src: genversion Only in amanda-2.4.2p2/common-src: genversion.h Only in amanda-2.4.2p2.orig/common-src: version.c Only in amanda-2.4.2p2.orig/common-src: versuff.c diff -ru amanda-2.4.2p2.orig/config/Makefile.in amanda-2.4.2p2/config/Makefile.in --- amanda-2.4.2p2.orig/config/Makefile.in Tue Apr 3 14:55:47 2001 +++ amanda-2.4.2p2/config/Makefile.in Tue Sep 17 18:02:16 2002 @@ -32,8 +32,6 @@ includedir = @includedir@ oldincludedir = /usr/include -DESTDIR = - pkgdatadir = $(datadir)/@PACKAGE@ pkglibdir = $(libdir)/@PACKAGE@ pkgincludedir = $(includedir)/@PACKAGE@ diff -ru amanda-2.4.2p2.orig/docs/Makefile.in amanda-2.4.2p2/docs/Makefile.in --- amanda-2.4.2p2.orig/docs/Makefile.inTue Apr 3 14:55:45 2001 +++ amanda-2.4.2p2/docs/Makefile.in Tue Sep 17 17:59:34 2002 @@ -32,8 +32,6 @@ includedir = @includedir@ oldincludedir = /usr/include -DESTDIR = - pkgdatadir = $(datadir)/@PACKAGE@ pkglibdir = $(libdir)/@PACKAGE@ pkgincludedir = $(includedir)/@PACKAGE@ diff -ru amanda-2.4.2p2.orig/example/Makefile.in amanda-2.4.2p2/example/Makefile.in --- amanda-2.4.2p2.orig/example/Makefile.in Tue Apr 3 14:55:45 2001 +++ amanda-2.4.2p2/example/Makefile.in Tue Sep 17 17:59:49 2002 @@ -34,8 +34,6 @@ includedir = @includedir@ oldincludedir = /usr/include -DESTDIR = - pkgdatadir = $(datadir)/@PACKAGE@ pkglibdir = $(libdir)/@PACKAGE@ pkgincludedir = $(includedir)/@PACKAGE@ diff -ru amanda-2.4.2p2.orig/man/Makefile.in amanda-2.4.2p2/man/Makefile.in --- amanda-2.4.2p2.orig/man/Makefile.in Tue Apr 3 14:55:45 2001 +++ amanda-2.4.2p2/man/Makefile.in Tue Sep 17 17:59:17 2002 @@ -34,8 +34,6 @@ includedir
Re: Amanda and ADIC tape library
On Tue, 17 Sep 2002, Alan Horn wrote: Is anyone currently running amanda on solaris with an ADIC Tape library ? Sorry, can't help you with the details. We have an ADIC Scalar 1000 here that I'd like to get working, any help appreciated. My big problems right now are with the output from mtx not being understood by amtape/amlabel etc... Are you using chg-zd-mtx or chg-mtx? Read TAPE.CHANGERS and make sure you've got the right mtx script for your version of mtx. If you're sure you're using the right mtx changer script then can you provide more details such as copy-and-pasted error messages from amtape/amlabel etc? Also, does anyone have a good tapetype definition for Sony SDX-500C AIT2 tape drives ? There is one in the FAQ-O-Matic at amanda.org IIRC. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: simple query
On Fri, 6 Sep 2002, Galen Johnson wrote: OK...I'm still reading as much documentation as I can find before I really decide I'm ready to tackle Amanda...One thing that keeps cropping up is that if Amanda hits the end of tape during a dump, it jumps to the next tape and starts over. This brings up the situation where the dump itself is greater than an empty tape. Let's say it crops up (worse case) at the beginning of a run...wouldn't this seem to be antithetical to running a backup? If followed to it's logical conclusion, you could potentially run through your entire tapeset without backing anything up, even though the other backups following the runaway backup might happily fit on the tapes. Or am I missing something? I realize this may not be a critical problem (just readjust your disklists the next day) but does seem possible (not to mention a nuisance). Amanda will not use more than the number of tapes you have set runtapes to in a single run so there's no way she's going to overwrite your whole tapecycle unless you're foolish enough to sit there and keep feeding her tapes. If you do indeed end up with a filesystem which can no longer fit on tape your best bet is to split it up via tar into smaller component directories. When you do this amanda will no longer care about the original large directory and you can safely remove it from your holding disk as soon as subsequent amdump runs get level 0s of the new subdirectories you have defined. If you're backing up user data and home directories you can also just enforce quotas on the directories at the outset and announce to your users the maximum ceiling you are able to backup. This is what we do and it has many side benefits as well. We provide each user with a seperate scratch directory they are free to fill to their heart's content which sits on a RAID5 but doesn't get backed up. They all know that this scratch space is relatively safe but is not guaranteed and it is certainly vulnerable to user error. They are all encouraged only to keep data which cannot be regenerated in their home directories and to regularly copy intermediate results of the analyses they run back to their home directories. Plus, we encourage archival runs of their data at the end of the project which will allow them to remove it from their home directory altogether. As they finish a project they have the option of sitting down at a machine we have made available which supports all manner of storage media from DDS to AIT, from CD-R(W) to zip disk, each according to his capacity, and moving the data offline. ;-) This is where it gets interesting for me. We are looking to replace this last step with some sort of large tape library and an HSM or hierarchical storage management solution. There do not currently appear to be any good open source software solutions for management of an HSM scheme. DMF from SGI/Cray is the only decent commercial package that does it properly and it's prohibitively expensive. I've begun to give a lot of thought to what modifications would be necessary to make such an open source application amanda derived. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Tape technology
On Fri, 6 Sep 2002, Gene Heskett wrote: On Friday 06 September 2002 13:24, Brandon D. Valentine wrote: s/They're/Their/ s/GB/TB/ You've been using vi too long, we can tell. And sed and perl too. [ This email written in vim, the one true editor. ;-) ] -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Tape technology
On Fri, 6 Sep 2002, Brian Jonnes wrote: I will be looking to get a new tape drive, but am not very familiar with the current technology. I have heard of DDS, DAT and have used Travan. As far as I'm concerned, the Travan is out of the question, 'cause the tapes are so expensive (and apparently they are now obsolete?). As Gene Heskett indicated elsewhere in the thread the appropriate tape technology depends on your budget and intended use. DAT/DDS is perfect for a home user. Far better than QIC/Travan ever was in terms of the price to performance ratio. However, I wouldn't currently sink any money into it as a new solution for commerical/business use. Sony has announced that DDS4 is the end of the road for DAT based backup technology. They're transitioning entirely to AIT. I've been using AIT for going on three years now and I've been extremely happy with it. It's more expensive than DAT formats, but it's getting cheaper. I've used AIT1 for the past three years and am currently transitioning to AIT3 technology. AIT1 is 35GB native capacity, AIT2 is 50GB and AIT3 is 100GB native capacity. With the release of AIT3 drives and tapes the prices on AIT1 are falling to where they can begin to compete with DDS4 -- which is limited to 20GB. Plus AIT drives can stream at or near the speed of a conventional hard disk in my experience. Amanda does an excellent job of keeping the drive busy via taper. DDS is a 4mm tape format and AIT is an 8mm tape format. Both of these are rather small form factor cartridges which make for easy storage. Both technologies achieve their capacity through helical scan technology and use double-reel cartridges. LTO is currently positioned to replace DLT in the high-end tape market. LTO stands for linear tape-open and the cartridges contain a single reel of tape which is wound onto a reel in the drive instead of remaining in the cartridge. Linear tape technologies like this achieve their density by cramming a whole buncha tape into a really big cartridge, leaving no room for a second reel. I get the willies at the thought of my tape being wound out of the cartridge and into the drive, but it seesm to work fairly reliably. It it's high-capacity format, Ultrium, LTO currently supports tapes of 100GB or 200GB native capacity. Personally at the prices they want for those suckers I'd just as soon buy a 100GB native capacity AIT3 drive with a fat stack of tapes. The AIT3 cartridges are small and easy to store and if you're planning to use amanda -- where you likely rewrite the tapes every couple of weeks -- it's hard to beat AIT's Advanced Metal Evaporate tape medium and the fact that you can actually afford to replace AIT tapes when, not if, you have one wear out. Also, Sony will release SAIT-1 towards the end of this year which is their own single-reel cartridge format -- though they're sticking with helical scan instead of linear tape technology. SAIT-1 will debut at 500GB native capacity in a cartridge whose size is equivalent to that of an LTO cartridge. They're roadmap takes them up to SAIT-4 at 4GB native capacity by 2010. Crazy man. Hope I've provided some food for thought, -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Trivial question - No Flames please
On Sun, 1 Sep 2002, Potts, Ross A. wrote: I think I have been looking in all the wrong places. What size is a 112 Meter 8mm tape with and without compression? I just inherited a butt-load of them and decided to use them for daily incrementals (nothing above 500 MB a day) Which 8mm format? AIT? Exabyte? More information is needed. Also, I'm willing to bet quite a bit that this is a fairly easy piece of information to locate yourself on Google. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: amdump and tar don't work
On Fri, 9 Aug 2002, thomas holstein wrote: FAILURE AND STRANGE DUMP SUMMARY: samson.hei /dev/da0s1e lev 0 FAILED [disk /dev/da0s1e offline on samson.heitec.net?] You must give tar directory names, not disk devices. Rather than /dev/da0s1e you need to tell tar to backup /var or whereever that partition happens to be mounted. Modify your disklist to reflect this. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
RE:
On Wed, 17 Jul 2002, Bort, Paul wrote: AMANDA is still actively being developed and used around the world. This is a good opportunity to remind everyone to take a few minutes to append their organization to the Who is using Amanda? question in the FAQ-O-Matic. http://amanda.sourceforge.net/cgi-bin/fom?file=319 -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
RE: My tape drive dissappeared after updating the operating syste m - now what?
On Mon, 15 Jul 2002, Martinez, Michael - CSREES/ISTM wrote: I use /dev/nst0 (no-rewind) in my amanda configuration, and it works fine. [snip] From: Bort, Paul [mailto:[EMAIL PROTECTED]] [snip] Also, I hope you're using /dev/nst0 instead of /dev/st0 in your amanda configuration, or your backups may be no good. [snip] Who's on first, What's on second and I Don't Know's on third. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Remote Tape drive
On Fri, 12 Jul 2002, Ashwin Kotian wrote: Your co-operation suggestions are very much appreciated. man rmt man rdump -- pay attention to '-f' argument man rrestore -- ditto man tar -- pay attention to --file argument, GNUtar only Lather, rinse, repeat. Now you have the information necessary to ask smart questions[0]. I don't know of any current support for this in amanda but it should not be difficult to implement. If you're intersted then look at tape-src/tapeio.[ch] in the amanda source. Sounds to me like implementing that is overkill for your project though. Just install amanda on the machine with the tape drive; run two concurrent amandas on that machine if it's already running an amanda install until you get your testing done. If you're really worried about it, dig up an old QIC drive or something to test with. [0] - http://www.tuxedo.org/~esr/faqs/smart-questions.html -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: who uses amanda?
On Tue, 9 Jul 2002, Ronald O. Christian wrote: The IS manager wants to be assured that Amanda has wide usage at large companies or organizations. I know the National Institute of Health and the US Department of Energy uses Amanda. Can anyone give me other names the IS manager would recognize? (Utilities would be best, but I will take anything.) Since this question comes up often I've added it as an FAQ entry under the Miscellaneous category. http://amanda.sourceforge.net/cgi-bin/fom?file=319 I think I've got it off to a good start. Go forth and enter information about how and where you're using Amanda. Just click on append to this entry to add your own testimonial. You'll have to register for an FAQ-O-Matic account if you don't already have one. Maybe if we get a good enough response the folks maintaining the amanda.org website will put up a direct link to the Who's using Amanda? entry. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Problems with amanda
On Wed, 3 Jul 2002, Robert Kearey wrote: That has not been the general experience of long time contributers to this list. It may then be time to re-evaluate that perception. Here are my thoughts on the matter: Package management is invaluable from a system administration standpoint; there's no question about that. However, vendor provided packages are something I deem inappropriate for an application as configuration intensive as amanda. I don't want to install a package and then have to tweak anything post-package install. If I am going to have to do that I might as well just NFS mount the source and run 'make install'. The better solution to managing your amanda install is to create your own packages. It's not terribly difficult to take the existing amanda RPM and convert it to suit your needs. It's also not terribly difficult to create custom apt packages for your Debian systems or create custom inst packages for your SGIs. This is the approach I prefer to take to managing my systems. For instance, we use OpenSSH on our SGIs but the SGI Freeware package for OpenSSH is almost always out of date and rarely do the included configuration files suit our needs. So, I maintain custom inst packages for internal use on our machines. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: amandates file
On Wed, 26 Jun 2002, Scott Inglis wrote: What is an amandates file and where do I get it? /etc/amandates is a file which must exist and be writable by the amanda user on each machine. You must simply create an empty one by typing touch /etc/amandates as root and then chown chgrp it to the amanda user and group. From there on out you'll be good to go. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Need advice on backing up a large directory
On 24 Jun 2002, Kirk Strauser wrote: I'm in the process of converting about 400 CDs to .oggs. The destination directory is currently approaching 9GB in size, and I'm using Amanda to backup to DDS-3 (12GB) tapes. Given that oggs and MP3s are essentially uncompressible, I'm getting pretty close to the limits of what I can fit onto one tape. Wow, that's dedication. I would never even think of backing up my MP3s. I've got over 500 CDs in my collection and they're all ripped but my inclination is just to throw them on a cheap IDE RAID. If something catastrophic happens and I lose the RAID, oh well, I've got the CDs and it didn't take me /that/ long to rip them. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
RE: Who uses Amanda?
On Mon, 24 Jun 2002, Martinez, Michael - CSREES/ISTM wrote: We use it: Research, Education and Extension Program Department of Agriculture Washington, DC -Original Message- From: Paul Bijnens [mailto:[EMAIL PROTECTED]] Sent: Tuesday, January 29, 2002 10:33 AM To: [EMAIL PROTECTED] Cc: Amanda Users Subject: Re: Who uses Amanda? On Fri, 25 Jan 2002, KEVIN ZEMBOWER wrote: - (This might seem like a stupid question to this group, but) I'm being - challenged by the folks who can't get my firewall setup to work with - Amanda that I should adopt a more industry-standard backup product. - Hogwash. But, I would like to at least offer an answer. - - Anyone have any guesses how many institutions and individuals are using - amanda? - - Anyone know, or want to self-disclose, some noteworthy institutions - using amanda? If you think this would clog up the list too badly, email - me privately at [EMAIL PROTECTED], and after a week or so, I'll - compile a list and post it to the email list. 1) You've fallen prey to the recent barrage of queued messages which were reinjected into the amanda-users mailing list. Note the date on the message you responded to is 'Tuesday, January 29, 2002 10:33 AM'. 2) Seeing this old message brought up a good point. It would probably help sysadmins pushing amanda adoption if amanda.org had some kind of Who's using Amanda? section featuring testimonials, similiar to what the Linux/XFS project has[0]. [0] - http://oss.sgi.com/projects/xfs/xfs_users.html -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: AIT libraries
On Mon, 22 Apr 2002, Bernhard R. Erdmann wrote: Once upon a time I ran an Overland Library with 19 AIT-2 tapes and a Sony SDX-500C internally just fine with Amanda. Hey Bernhard, I appreciate your response. I am very interested in the Overland Data products now. Especially after hearing of Stephen Carville's troubles with the Spectra Logic Treefrog LVD SCSI adapters. The Overland Data products seem to be price competitive with the Sony offerings and include 19 as opposed to 16 slots, which gives me much more breathing room in the library. I like that number a lot better. Anyhow, my question for you, what changer glue were you using with this library (i.e. chg-scsi or something else) and did you use a barcode reader in your setup at all? That would be good information to have before I recommend one of these for our setup. Thanks, -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
floodgates
If it's at all possible, stop the insanity! Some SMTP server in Germnay has opened its floodgates. Is it possible for someone with access to the machine running the amanda listserv to deny relaying from: mailto.t-online-com.de mail.t-intra.de Thanks, -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: your listserver?
On Sun, 23 Jun 2002, Gene Heskett wrote: Unforch, as a subscriber to this list, my only choice is to pop filter mailto.t-online-com.de, flagging any header that contains this address as an auto delete without ever downloading the message. FYI, this has been working fine for me out of procmail: :0: * ^Received:.*mailto.t-online-com.de amspam I could also have /dev/nulled them but I'm interested in counting them when the noise dies down. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Tape changers/libraries
On Fri, 3 May 2002, David Flood wrote: I have around £2000 to spend on a changer/library does any one have any suggestions, ideally I would like to be able to back up 400GB but I realise this is maybe not possible for 2K so backup capability is the most important factor. Hi David. I've been doing a lot of research lately evaluating various tape library/changer products for our use here. I could share some of my insights so far with you, but it would help to have an approximate conversion of £2000 into dollars. I don't follow currency exchange very closely. ;-) -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: ext2 and mail (Was: Backing up PostgreSQL?)
On Fri, 21 Jun 2002, [iso-8859-1] Martín Marqués wrote: On Jue 20 Jun 2002 17:17, Ragnar Kjørstad wrote: In particular there is a thread about filesystem-features required for mailservers, and there is a post from Wietse where he writes: ext2fs isn't a great file system for mail. Queue files are short-lived, so mail delivery involves are a lot of directory operations. ext2fs has ASYNCHRONOUS DIRECTORY updates and can lose a file even when fsync() on the file succeeds. Sorry, but I think Wietse was talking about ext3, not ext2. ext3 has asyncronous updates, as it's a journal FS. Not an expert on FS, but I read Wietse's mail on the Postfix mailling list. Negative. ext2 by default on every linux distro I've ever encountered is mounted async. You /can/ mount ext2 filesystems synchronously but performance falls through the floor, the earth's crust and well into the mantle. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Please help me!!
On Tue, 18 Jun 2002, Steve Bertrand wrote: I have been working at this for 14 hours now, and still can't get it. I have even wiped amanda, reinstalled, then wiped BSD and reinstalled it and amanda. Process of elimination: have you tried switching your dumptype for this disk to GNUtar to see if that fares any better? -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Backing up PostgreSQL?
On 13 Jun 2002, Kirk Strauser wrote: I just finished reading `Unix Backup and Recovery' and realized that there are quite a few holes in my disaster recovery plan. In particular, I'm looking for a way to backup PostgreSQL without taking the database offline. I'm assuming that whatever method I use will involve writing a wrapper script to prepare the backup, launch Amanda, then clean up afterward, but that's about as much as I've guessed right now. Checkout pg_dump(1). Setup a cron job which runs pg_dump to a logrotate style directory of database dumps which are picked up by amanda in its nightly run. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Backing up PostgreSQL?
On Thu, 13 Jun 2002, [iso-8859-1] Ragnar Kjørstad wrote: Alternatively to the pg_dump approach others have suggested you can snapshot your filesystem and then use regular backup. The easiest way to snapshot the filesystem is to use a logical volume manager (LVM or EVMS on linux) and then do: 1. take database offline 2. take snapshot 3. take database online 4. backup from snapshot 5. remove snapshot I would like to comment that while this is a possible way to backup your database, it's not the way I would recommend going about it. There are a couple of caveats: 1) In order to take a filesystem snapshot you must have enough diskspace elsewhere to contain the filesystem snapshot. If your database resides on a large filesystem with other data[0] then you're unlikely to want to deal with caching an entire snapshot until amanda backs it up. 2) Backing up a database by grabbing the actual files off of the disk can introduce problems if trying to restore onto, for instance, an upgraded version of Postgres, which might have changed the ondisk representation slightly. There are also problems if you migrate to a different architecture since things like byte ordering can change. By generating a database dump with pg_dump you insure that you have a portable, plain text file full of valid query commands which can be read into any future version of Postgres and possibly even into other RDMBS products provided you choose a pg_dump format which is standards complaint enough. 3) If your snapshot code is not atomic it means you must take your database server down everytime you make the snapshot, which on a large filesystem could be a non-trivial amount of time. With pg_dump you're just dumping the tables via the standard Postgres interface so you've got no issues with doing it on a running database. [0] - which is a very likely scenario since I know few people who make a seperate filesystem just to hold their database. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: LART
On Thu, 13 Jun 2002, Chris Dahn wrote: From the Jargon Lexicon: http://www.tuxedo.org/~esr/jargon/html/entry/LART.html Luser Attitude Readjustment Tool. The LART classic is a 2x4 or other large billet of wood usable as a club, to be applied upside the head of spammers and other people who cause sysadmins more grief than just naturally goes with the job. In our department I've dubbed our LART the Clue-by-four. Also acceptable are the Cluebat (when referring to a Louisville Slugger) and the Cluebar (useful for prying the top off of the Luser to examine the extent to which his brains are missing). -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Backing up PostgreSQL?
On 13 Jun 2002, Kirk Strauser wrote: Space permitting, I'd think that making an (ASCII) copy of the database contents into a file is a pretty smart idea, even if you're backing up the DB files in another manner. Space permitting? Most natural language plain text files compress very well. A file full of database queries is going to have lots of easy to predict repetitive query statements and lots of natural language within the database entries itself. It should compress better than 2:1. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: multiple drives - one config
On Sat, 25 May 2002, fil krohnengold wrote: Is there a way to utilize multiple drives in a single tape changer unit without splitting up the configuration? If you would like to use them in parallel then no, there is no way to do it without splitting the configuration between the drives. You could in theory use chg-multi to use your drives serially but I see little point in that if all of your tapes are in the same changer since you're still only using one drive at a time. It seems to me there is little reason to own more than one tape drive with amanda unless you've got multiple configs going or you keep an extra drive around as a spare. If any of the amanda hackers have any better ideas I'd love to hear them. I've often wondered how amanda could be made to cope with this situation. It can't be that uncommon. Let's say you've got some 19 tape Overland Data AIT library, and you're running out of capacity in it, so you buy another identical library which requires using a different drive to access those 19 tapes. Seems a shame not to be able to specify runtapes 2 and have amanda use both drives in parallel. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
RE: cygwin/dump
On Thu, 23 May 2002, Bort, Paul wrote: 6. The only backup program I've ever seen that could restore a Windows NT box to a usable state is Ghost. If you really need fast restores, consider Ghosting the machine to a file and backing up that file. Ghost costs dinero. Check out Mondo Rescue: http://www.mondorescue.com/ Hugo (the Mondo developer) did a demo for us here where he brought in his actual personal workstation, used Mondo to back it up onto CD (it also works with tape), and typed rm -rf /. Naturally the crowd went wild over actually seeing someone purposefully type that command. But he let it run until the OS destroyed itself, then he just popped in the first CD in his restore set, hit the reset switch and within 30 minutes he had his entire disk back exactly as it was. He eats his own dogfood in this manner several times a week. Mondo is well tested. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: making 2.4.3p3 on octane running 6.5.12
On Wed, 22 May 2002, Bernhard R. Erdmann wrote: Chris Stone wrote: I'm having trouble building the amanda client on an sgi. I get the same error on a SGI Indigo2 running IRIX 6.5.15 and using gcc 3.0.4 and gmake 3.79.1 from freeware.sgi.com: To add a datum to the context, I have no trouble building and installing amanda on IRIX with the MIPSpro toolchain. I have not ever had occasion to try the gcc route. If you access to MIPSpro compilers you might try them. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: 24 gig Tape Settings
On Sat, 18 May 2002, Alexander Belik wrote: filemark 0 kbytes What is this? AIUI this is an approximation of the amount of space a filemark occupies on tape. Filemarks seperate files on tape, and because amanda stores each disk in the disklist in a seperate file on tape, it must account for any loss of capacity due to filemarks when estimating. IIRC though this value is not currently used by or necessary for amanda so it's perfectly safe to set it to zero which is what most of the tapetype entries you will find in the FAQ-O-Matic and here on the list have set it to. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Changing From 8gig tapes to 24 gig tapes
On Fri, 17 May 2002, Brook Hurd wrote: I am about to change my 20 tape backup process from 8 gig tapes to 24 gig tapes. I would appreciate anyone's insite on this process. Any insites, warnings, or instructions would be greatly appreciated. I would create a temporary testing configuration to make sure your tapetype for those 24gig tapes works right. You can give all the new 24gig tapes identical labels to your old 8gig tapes with 'amlabel -f'. Then I'd probably wait until the end of a tapecycle and just flip tapetype over in the amanda.conf. Planner should automatically start recalculating its runs based on the increased tape size and amdump should accept the 24gig tapes provided they have the same labels. Each day amanda will think it is overwriting one of your old 8gig tapes. By the end of the tapecycle you should be in good shape. You absolutely must keep the 8gig tapes around until at least the end of your tapecycle though in case you need to restore off of them. Restoring off of a mix of 8 and 24 gig tapes should not cause a problem provided you know which 8 gig tapes have been replaced and which are still in the cycle. This is the main reason for waiting for the end up of a tapecycle before making the switch, it's easier to mentally keep track of what's been replaced so far if it correlates numerically. HTH, -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: 24 gig Tape Settings
On Fri, 17 May 2002, Brook Hurd wrote:
I am about to use the following tape type:
HP C5708A DDS-3 Data Cartridge, 24GB
Unfortunatly, I don't see it listed online. Anybody use this type
before, and if so, what are the correct settings?
This is a 12GB native capacity DDS3 cartridge. Are these new tapes or
have they been used before? You will have to decide whether you want to
use hardware or software compression. With software compression you
tell amanda the exact native capacity of the tape. If you use hardware
compression you will have to guess about how well your data compresses
and lie to amanda about your tape capacity. I prefer to use software
compression and I would recommend that unless you're an amanda expert
you stick with software compression if you can. If these tapes have
ever been used before in a DAT drive with compression enabled you may
have trouble getting them to work with hardware compression disabled.
There was a thread here recently about degaussing DAT tapes to erase the
compression bit. You might look for it in the archives. If you don't
use hardware compression there are a number of DDS-3 tapetypes in the
FAQ-O-Matic which will work, depending on what model of DDS drive you
are using. Here's one for a Seagate drive for example:
define tapetype SEAGATE-DDS3 {
comment Seagate STD224000N-SB Internal DDS-3 Drive
length 11550 mbytes
filemark 0 kbytes
speed 1075 kps
}
Good luck.
--
Brandon D. Valentine [EMAIL PROTECTED]
Computer Geek, Center for Structural Biology
This isn't rocket science -- but it _is_ computer science.
- Terry Lambert on [EMAIL PROTECTED]
Re: amanda and Procom's NetForce-800d
[ sending one copy of the message will suffice. thanks. ] On Thu, 9 May 2002, Ted Sariyski wrote: I have been using amanda 2.4.1p1 more than two years for a couple of dozens of Unix machines and it works great. Now I add a NetForce-800 file server. The only way to access the file system of this server is through nfs or smb. For now I have no option to install a client directly on the server because it will violate the warranty. The file server support checkpoints and has it's own proprietary backup system but I would like to keep the backup uniform over the network. Until now I have no reason to replace amanda with a commercial solution. Is there a way to force amanda to dump a file system over nfs? This is an easy problem to solve. I have a machine which I could not get a reliable amanda client working on (due to some Linux library madness I don't even want to start into it). Until I am able to replace that machine I mounted it via NFS on my backup server and put the disks into disklist as if the filesystems were located on the backup server in the places I had mounted them (actually they were automounted). It has been working in that configuration flawlessly for six months and I have even restored several things to that machine in that time. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Tape changers/libraries
On Fri, 3 May 2002, David Flood wrote: I have around £2000 to spend on a changer/library does any one have any suggestions, ideally I would like to be able to back up 400GB but I realise this is maybe not possible for 2K so backup capability is the most important factor. Hi David. I've been doing a lot of research lately evaluating various tape library/changer products for our use here. I could share some of my insights so far with you, but it would help to have an approximate conversion of £2000 into dollars. I don't follow currency exchange very closely. ;-) -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Using Multiple Tape Sizes
On Thu, 2 May 2002, Don Potter wrote: I would be willing to bet that you could use..them but since the config specifies the tape size, you would only be using the limits of the previous size. I would say add the tapes in your rotation and adjust your tape size to reflect the new tapes. And remove the the 8GB tapes from your rotation. As Don says, you're limited to the size of the lowest capacity tape in the cycle. But if you've got enough tapes to replace the entire cycle this procedure might work well: 1) Use 'amlabel -f' to give the 20GB tapes the same labels as the old 8GB ones. 2) Switch tapetype in amanda.conf over to a 20GB capacity. 3) Start the next day by backing up to the tape amanda asks for but from the 20GB instead of 8GB tapes. 4) Keep your 8GB tapes around until the end of the backup cycle. Should you be asked in the middle of the cycle to restore something you'll need to have the previous cycle's worth of tapes to feed to amrecover. 5) You should have succesfully slipped a new tape cycle in. What you want to avoid is using 'amrmtape' on your 8GB tapes because as soon as you do that you'll no longer be able to restore off of them. If you don't yet have enough 20GB tapes to replace your entire 8GB cycle you might want to think about doing the above but without step 2. You can use just the first 8GB of the 20GB tape without any trouble. Then if you later on got the money to replace the rest of the 8GB tapes you could cycle them in and switch over to a 20GB tapetype. Anyone out there see a problem with this theory? -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Backing up advice
On Mon, 29 Apr 2002, Chris Mason wrote: Can anyone point me at some information on setting up a backup schedule, how to organize tape rotation, etc. I'm planning a tape backup system for a small business and need to write a HOWTO for them. I'm sure it would be convenient if we did the typing for you, but we're kind of busy. UTFW. amanda.org has links to all of the documentation yet compiled on amanda. There are people in the Appalachian Mountains who've never worn shoes who can use Google. Try it. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: AIT libraries
On Mon, 22 Apr 2002, Bernhard R. Erdmann wrote: Once upon a time I ran an Overland Library with 19 AIT-2 tapes and a Sony SDX-500C internally just fine with Amanda. Hey Bernhard, I appreciate your response. I am very interested in the Overland Data products now. Especially after hearing of Stephen Carville's troubles with the Spectra Logic Treefrog LVD SCSI adapters. The Overland Data products seem to be price competitive with the Sony offerings and include 19 as opposed to 16 slots, which gives me much more breathing room in the library. I like that number a lot better. Anyhow, my question for you, what changer glue were you using with this library (i.e. chg-scsi or something else) and did you use a barcode reader in your setup at all? That would be good information to have before I recommend one of these for our setup. Thanks, -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
amrecover weirdness
This is not a problem, per se, but it Should Not Happen . When I launch amrecover from a directory which is in the disklist I get this: [root@moby admin]# amrecover AMRECOVER Version 2.4.2p2. Contacting server on backup ... 220 moby AMANDA index server (2.4.2p2) ready. 200 Access OK Setting restore date to today (2002-04-25) 200 Working date set to 2002-04-25. 200 Config set to Daily. 200 Dump host set to moby. Can't determine disk and mount point from $CWD amrecover lpwd /export/moby/admin I can't use amrecover until I run setdisk /export/moby/admin. After that it works like a million bucks. It's been a while since I ran a restore so I had forgotten about this behavior, but I just was reminded of how annoying it is. What can be done about amrecover not knowing what the CWD is? -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
AIT libraries
This is exciting. My department is seriously looking at investing in a tape library. I've always wanted one to call my own ;-). So it is with this happy news that I turn to you, the most helpful group of tape library owners I am aware of. I am seeking recommendations, testimonials if you will, from people succesfully using AIT libraries with amanda. To aid in this, here's a bit of background about our current setup and our projected budget storage goals. We currently use a standalone SDX-300C AIT1 tape drive w/ 35GB tapes. We run a 2week long 10tape Daily backup cycle and a Monthly 10tape cycle. The Monthly cycle uses skip-incr, record no and chg-manual. However, we are rapidly accumulating data beyond what I can realistically do a chg-manual, skip-incr dump on. I am interested in AIT libraries in the 15-16 tape size range. Solid compatibility with amanda is a must. We are looking to spend obviously as little as possible to get a quality product, but $10k is a very firm ceiling and the closer to $5k we can hold the price, the more likely it is to get funded. Rackmount is a must, though I'm not so concerned with how many RUs it takes up. So far the only products I am really familiar with are the Sony LIB-152 LIB-162 series. I understand that Overland Data and several other companies are also making AIT libraries. I'd be interested to hear about those. Pros/Cons of them versus the Sony ones, etc. My other question is: Does anyone know how backwards compatible AIT2/AIT3 drives are with AIT1 tapes? The reason I ask is that if the backwards compatibility is good I would consider the extra cost of an AIT2 or 3 drive in the library up front so that if (and when) the capacity of our existing tapes started to become strained I could spring for the relatively cheap upgrade of a new round of larger tapes. Those 100GB native capacity AIT3 tapes are quite sexy. Thanks, -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
tape server crash recovery
So early Saturday morning our tape server suffered a kernel panic, and it appears to have happened smack in the middle of the amdump run. When I brought the machine back up I looked and found that the run had not completed, so I ran amcleanup, which appears to have taken care of the logfiles left behind. Then I looked and found a directory still in the holding disk from that night, which contained 4 incomplete dumps left behind by the 4 active dumpers at crash time. I removed this directory. Then I ran amcheck to see if anything else was odd. Where it should have told me which tape was needed for the next run, it only said that it was expecting a new tape. I went and looked at the tapelist file for this config and it was empty! Amdump must have been accessing it in some way when the crash occured. Luckily tapelist.yesterday had not been overwritten so I looked at it and found that it was exactly what I needed: 20020329 Daily03 reuse 20020328 Daily02 reuse 20020327 Daily01 reuse 20020326 Daily10 reuse 20020323 Daily09 reuse 20020322 Daily08 reuse 20020321 Daily07 reuse 20020320 Daily06 reuse 20020319 Daily05 reuse 20020316 Daily04 reuse The 20020330 run (Saturday morning) was set to reuse Daily04, and that was what was in the drive when the crash occured. I copied tapelist.yesterday to tapelist, and ran amcheck again, which told me what I expected to see: Tape Daily04 label ok I started up amdump by hand (although it didn't background itself like one would think it would given all the other amanda programs do, I had to hit ^Z and run 'bg' from the shell to get my terminal back). It's running now and appears to be going well except that some of the disks are listed in the amstatus output like this: [dumps too big, but cannot incremental dump new disk] This seems to be contingent on some variable that is currently escaping me though. The amstatus output lists plenty of disks that have already dumped taped level 1 and level 2 incrementals since I started up amdump. Amdump doesn't seem to be ignorant of the run history, so I wonder why it is singling out these certain volumes as 'new disks'. Perhaps these are the volumes in the process of being dumped or taped when the crash occured? Thoughts? Suggestions? If I force full dumps of the afflicted partitions for tonight should I be safe? -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: amflush and irc?
On Thu, 21 Mar 2002, Jon LaBadie wrote: You won't. After running a while amanda will spread the level 0's over all the dump days. So if your once a week tape now gets 40GB and your daily incrementals now get 2, each will average about 8 GB daily. Actually it's even better than that. Each tape should average a full 40GB. Amanda will do its best to fill each and every tape to capacity, promoting as many full dumps per run as possible. If you have 40GB of data and a 40GB tape drive, you should get a full set of level 0s on every run. If you have 60GB of data and a 40GB tape drive you should see a full dump of everything every 2 days, with incrementals in between, etc. It's really cool. Some might suggest that if you can't justify the expense of tapes then the value/cost of losing the data must be low. That's a valid point to consider. Also worth noting is that if you are concerned about raw $/GB in terms of tape you can't go wrong with DAT-based formats. They may not be the most advanced, robust tape technologies out there, but DDS ain't bad and the media sure is cheap. We use AIT at work but I use DDS3 for all of my personal backups at home (I also use amdump as my alarm clock each morning, as soon as that thing starts taping I find I can no longer sleep ;-). -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Inconsistent taping
On Wed, 6 Mar 2002, Eric Trager wrote: Is what hardware compression can do with files variable? In this case, with the failure, it's a lot of oracle data, some of it is gzipped. This is precisely why I do not use hardware compression. The effectiveness of the compression varies heavily based on what kind of data you are backing up. Chances are 50GB of gzip'd data is not going to compress any at all. I run a fast dual PIII as my tape server and it handles doing the software compression for me just fine. With software compression planner's estimates are much more accurate and I sleep easier at night knowing that my filesystems are reliably getting taped, and if there is a problem that anything that wasn't taped was compressed and stuck in my holding disk to be flushed the following morning. My holding disk probably wouldn't hold an entire run uncompressed, but with software compression a $100 hard disk can hold an entire run. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: OnStream DI-30, amrecover not working
On Sun, 3 Mar 2002, Ray Shaw wrote: On the other hand, if they really are terrible drives, could I get some recommendations? I'd like something about that size (15GB uncompressed), and I could go with either internal IDE or external SCSI (probably better for various reasons). The idea is to get something we can actually buy...I'm not worried about winning any speed competitions, I'd just like to have money left over to fix the printer, and have tapes that will last for a while :) DDS3 drives are ~$250 dollars on pricewatch.com right now. The DDS3 DAT tapes are under $10 for 12GB native/24GB compressed. AIT1 and DDS4 drives run about the same, ~$550 for a decent drive. The tapes for DDS4 are under $20 for 20/40GB. AIT1 tapes are ~$50 for 25/50GB and ~$60 for 35/70GB. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Merry Christmas and a happy New Year
On Mon, 25 Feb 2002, James wrote: Thus spake Bill Sidhipong ([EMAIL PROTECTED]): It could just be misconfigured mail server :-) Seems innocuous enough, and though I've seen a marked increase in junk mail I receive (is any spammer harvesting amanda-* lists?) I've not seen the original message any more. Yesh. I'm willing to bet that's the case. amanda-users-list is probably the resend alias for amanda-users. It probably hasn't been secured, so anyone can send to it and bypass all the majordomo and resend criteria. Indeed and it pisses the hell out of me. It's not so freakin' hard to secure majordomo. People keep using the damn resend alias instead of the list address and I've had to add an additional lameness filter to my procmail setup. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Merry Christmas and a happy New Year
On Mon, 18 Feb 2002, Lewis Watson wrote: Yup. Has been delivered here about six times, we scan for virus' at the server and its not seeing anything. Nothing at sarc.com at this time... Regardless of whether or not it's a virus can people PLEASE stop replying to the amanda-users-list address? The correct address to send mail to this list is [EMAIL PROTECTED] I do not know why these virus/spam emails get sent there, but when you reply please change the header. If you send them to amanda-users-list you bypass some majordomo aliases which insert useful headers. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Hmm. Ok
On Mon, 18 Feb 2002, Lewis Watson wrote: Hmm. Ok. The only reason I was responding here was because they came from within the list. Sorry to have inconvenienced you. I don't think you understood me. It's perfectly okay to reply *here*, this is [EMAIL PROTECTED] If you will all go look back through your INBOX you'll find in the headers that the offending emails were To: [EMAIL PROTECTED]. Emails sent to that address bypass part of the majordomo resend mechanism. If you compare the headers between this message I'm sending you via the list and the spam message you'll see the difference. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Problem backing up SNAP server via Samba
On Fri, 15 Feb 2002, Dave Edick wrote: I'm in the process of setting up an Amanda backup server to back up a collection of new servers. The Linux and Solaris servers aren't posing any problems so far. But I also have to back up a SNAP 4100 appliance. In my old pre-amanda backup scheme, I was backing it up as an NFS volume, but evidentally Amanda doesn't do NFS volumes (please correct me if I wrong on this). So I set it up to back up using Samba. I installed the latest Samba and ran a manual smbtar, which worked. But when I run the nightly Amanda backup, it wigs out. It shows that it's backing up about 3.5Mb, but there is really about 3GB on the SNAP server. I included the output from amdump below. I run full backups every night. Any ideas what I'm doing wrong? In my opinion, adding the complexity of samba is wholly unnecessary unless you are actually backing up a Windows system. I have one linux machine here which has exhibited some very bizarre glibc issues and so I am unable to get an amanda client working on it. This machine at this time is a temporary situation and will get replaced in the next couple of months. In order to facilitate backing it up in the meantime I have it NFS exporting itself to my amanda backup server. The amanda backup server then automounts those partitions from the client. In the disklist I just have: backuphost /path/to/nfs/mount tar-NOTdump It's very straightforward. Amanda will backup any mounted partition, including NFS ones, so long as you are using GNUtar. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Tape changer question
On Fri, 15 Feb 2002, Ian Eure wrote: Anyways... can you divulge how I'd get this running, or suggest a better solution? It seems silly to stick 6 DAT tapes (24gb) in a tape cartridge that has to be swapped out every day when the daily incremental backups aren't breaking 250mb. I'm completely open to alternate solutions if you can suggest one. Umm, excuse me if I'm being rude, but duh. You stick 6 DAT tapes in the cartridge on Monday, and let Amanda use /one/ tape each night. It automatically gets a full set of level 0 dumps over the course of the week, as well as incrementals every night. This is what amanda is designed to do and it works damned well. I use 10 35GB AIT tapes per two weeks to backup my data and I see amanda getting a full set of level 0s done every three days or so, plus incrementals every night. Amanda is not a traditional backup system where you get full dumps weekly and incrementals every night. Amanda is designed to minimize time between full dumps, so you can get them more than once a week, and to insure that you always have incrementals to restore to any day within the dumpcycle. It is also designed to fill up your tapes. You should never be putting 250mb on a 24gb tape. You should always be putting 24gb on the tape. Amanda will do this for you. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: 2.4.2p2 on Irix 6.5.x, user=amanda, group=sys
On Mon, 4 Feb 2002, Brian Cuttler wrote: Unless, IRIX, unlike Solaris, is stripping the suid bit across the NFS network mount. I mean, this works fine on the Solaris config but is a problem on IRIX. IRIX, like unix only different. Come on, be fair. IRIX, like _ANY_ unix will strip the suid bit if you specify that behavior, which sounds like the case here. Read fstab(4) under IRIX for information on the nosuid mount option. If you don't desire this security-enhancing behavior just remove the nosuid option from your NFS mount. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: 2.4.2p2 on Irix 6.5.x, user=amanda, group=sys
On Mon, 4 Feb 2002, Brian Cuttler wrote: Should I expect this error when running amcheck on the amanda server which also have /usr/local as a local XFS partition ? I use NFS automounted amanda clients here with no problems. I have the partition with the amanda software installed exported like this(although I have genericized this some): --- /etc/exports /export/apps@ournetgroup(rw) --- --- auto.master /mnt auto.apps -intr --- --- auto.apps apps ourhost:/export/apps --- So that the amanda binaries would be in /mnt/apps/IRIX64/bin and /mnt/apps/Linux/bin respectively (so that the third level can be filled in by `uname -s`). Obviously the amanda UID/GID needs to be consistent across all clients whether it's maintained via NIS or by hand on each system. Otherwise the permissions on the binaries won't be right. I compile the clients on the NFS server hosting the apps directory with PREFIX set to /export/apps and then su to root in order to run the 'make install'. 'make install' takes care of all of the permissions including the various suid binaries. It works great for me. The Linux and IRIX clients have no problem using their binaries. The server is also a client and works great. I don't know what to tell you about the error you're seeing other than perhaps you missed a step somewhere in there. I can only hope reading this little writeup on how I've setup my NFS clients will remind you of something you forgot. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Simultaneous Multi-Tape Multi-Host Backups
On Wed, 30 Jan 2002, Aaron Rainwater - TXDC SysAdmin wrote: Okay, so if I go with the 4-Client-4-FileSys and 1-Host-4-TapeDrive situation. Can Amanda backup the 4 filesystems to the 4 tape-drives simultaneously from one config file? From what I've configured so far (with the chg-multi.conf), it can only treat drives as a disk-changer writing to them serially (even though they are 4 separate drives that could be written to simultaniously). I don't believe it can do it simulataneously, no. It can do it serially, yes. One possible way to get it going in parallel is to define 4 seperate config files, one for each client. Then if your cron job runs 4 copies of amdump, each with a different config specified, it should run in parallel. Each config just has a disklist which knows about the disks on its assigned machine. This might even have nice administrative side effects, making it easier to keep each machine on a specific set of tapes and being able to modify the configs on a per machine basis. Let the list know if you get it setup and running. I think we'd be interested to hear your success story. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Simultaneous Multi-Tape Multi-Host Backups
On Tue, 29 Jan 2002, Aaron Rainwater - TXDC SysAdmin wrote: Has anyone out there configured Amanda to make backups to multiple hosts and tape-drives simultaneously without having to create copious amounts of config files. The full realization of this concept would be to be able to backup data from, say, 4 filesystems on 4 machines to 4 tape-drives on 4 machines, simultaneously. Well, here's a theory. I have no idea if this would work, so YMMV. Setup Amanda in the usual fashion, i.e. choose a machine to run the Amanda server and install the client on all 4 machines. Configure the Amanda server with 4 seperate configs, one for each machine you want to backup. In each config find a way to make amanda use the remote tape drives via rmt (I don't know if this is possible, I'm sure one of the Amanda hackers could answer it). That might work. I however strongly question your motivations for wanting to setup something like this. I assume you want to do it because you already own 4 tape drives and want to make use of all of them? Here's a better suggestion: Find an obsolete machine you can devote to running Amanda and move all 4 tape drives into it. I have seen a 486 DX/2 66 w/ an Adaptec 1542 ISA SCSI card in it do DAE across 6 SCSI CDROM drives and not saturate the bus. You'll have no problem using an old 486 or Pentium class machine with any supported SCSI or IDE (depending on what kind of drives you have) controller card. 4 tape drives definitely won't stream fast enough to saturate the bus. I'd use something like NetBSD to keep the machine lean. Obviously the slower your tape server the more inclined you should be to use client compression, but the devil's in the details. You might even be able to write a changer script that would allow you to treat all 4 drives like a 4-tape changer and use just one amanda config. You'd be able to go a whole week without going into your closet to change a tape. I actually have considered doing this at the house with a stack of old Travan drives I have, for my personal backups. Load them up at the beginning of the week and forget about them. -- Brandon D. Valentine [EMAIL PROTECTED] Computer Geek, Center for Structural Biology This isn't rocket science -- but it _is_ computer science. - Terry Lambert on [EMAIL PROTECTED]
Re: Tape eject
On Wed, 16 Jan 2002, Stephen Hillier wrote: here's my crontab entry: 0 1 * * 1-5 /usr/bin/amdump DumpSet1 /usr/bin/amcleanup DumpSet1 eject /dev/st0 The call to amcleanup is unnecessary. From the amcleanup manpage: DESCRIPTION Amcleanup generates the Amanda Mail Report and updates the Amanda databases after a system failure on a tape server host. This cleanup process is normally done automatically as part of the amdump program, but if amdump cannot com- plete for some reason (usually because of a tape server host crash), amcleanup must be run some time later (usu- ally during system boot). Brandon D. Valentine -- Iam mens praetrepidans avet vagari. - G. Valerius Catullus, Carmina, XLVI
Re: smbclient question
On Mon, 14 Jan 2002, R. Bradley Tilley wrote: I have about 60 MS clients, and currently, they are all mounted by smbclient through the AMANDA backup server. I have a few other Linux machines that are backed up too, and I was wondering if it would be better to spread these SAMBA mounts out across 2 or 3 other Linux machines. So, instead of mounting 60 MS clients on the Amanda server, spread the 60 out across 3 Linux machines. Or does it matter? Any advice is appreciated. Seems rather Rube Goldberg to me. Why add an extra level of complexity for yourself when trying to debug problems? Brandon D. Valentine -- Iam mens praetrepidans avet vagari. - G. Valerius Catullus, Carmina, XLVI
Re: XFS, Linux, and Amanda
On Mon, 14 Jan 2002, Joshua Baker-LePain wrote: Hmmm, you did 'rpm -e' the RPM version, right? Pre-build amanda=bad. Word. Especially the moronic way in which RedHat has decided to build it. Brandon D. Valentine -- Iam mens praetrepidans avet vagari. - G. Valerius Catullus, Carmina, XLVI
Re: Dump to holding area only
On Mon, 7 Jan 2002, Don Potter wrote: I have two file systems that will no dump to holding_area statin g that they are too large: chgsfs /home5 lev 0 FAILED [dump larger than tape, but cannot incremental dump new disk] exon /export/home/bioproj00 lev 0 FAILED [dump larger than tape, but cannot incremental dump new disk] the filesystem derived from exon is about 10GB so I can't rationalize why the cojmplaint about lack of space when that isn't really the case. Anybody have any insight as to why this would fail?? If you have just started the cycle these messages will persist until all disks have received their first level 0 dump. It's expected behavior. You cannot incremental dump something which has no level 0 on tape. Thus, amanda does level 0s of as many filesystems as possible per day until everything is on tape. Give it some time to get through your disklist. Brandon D. Valentine -- Iam mens praetrepidans avet vagari. - G. Valerius Catullus, Carmina, XLVI
Re: tapetype recommendations
On Thu, 6 Dec 2001, Paul Bijnens wrote: Brandon D. Valentine wrote: My quarrel seems to be with the fact that amanda is attempting to backup more data than will fit on a single tape. This might not be a tapetype problem, but it seems the logical place to start twiddling. It may be that planner is unable to get good estimates for some of this data so what it thought would compress to 35GB only became 46GB post dump. That's probably the case. Planner gets an estimate of the uncompressed sizes, and makes an educated guess to what ratio it will compress. For it's education it uses the history of the last three dumps and computes an average. If it has no history, it assumes by default 50% (but you can change that in your dumptype with the comprate parameter). This seems to be the case as last night's amdump run also ran over the tape length, but by a noticeably smaller amount. I went ahead and set comprate to 0.90 in my global dumptype so hopefully tonight's run will be a bit better. If it's still not getting good estimates I will just continue flushing the leftover dumps to tape until planner figures things out. If for some reason, the data does not compress as well as it did in the past (e.g. because someone replaced an almost empty database with the same amount of mp3 files -- it happened to me) dumper has to consume a lot more data than planner thought it would. It's more like I finally got all of the bugs worked out of my heterogenous amanda setup which means everything in the disklist is finally getting dumped, and most are starting at level 0. I've set length in tapetype to 33000mb, a good 2GB below what the tape will theoretically hold. Worst case scenario is that Amanda uses powers of 2 to define tape size and Sony uses powers of 10. If this is the case then 33000*1024*1024/1000/1000 = 34603 so I should still be safe. Actually, amanda informs you about the exact place where it bumped into EOT: taper: tape Daily06 kb 34136032 fm 42 writing file: No space left on device That is 34136032 / 1024 = 5 Mb (Or should i count (34136032 + 42 * 500) / 1024 = 33356 Mb So you're indeed on the safe side with 33000 mb capacity in your tapetype. That was my hunch. Thanks for checking my sanity. =) Brandon D. Valentine -- Iam mens praetrepidans avet vagari. - G. Valerius Catullus, Carmina, XLVI
Static Compile
Alright. I'm currently going through a bit of fun trying to get a new version of the Linux amanda client compiled. I need to get it compiled statically since my clients are running various versions of the Linux system libraries. I was able to compile a static GNUtar just by setting the CFLAGS to -O -pipe -static before running configure and make. However, when I run amanda's configure and make with these options set I still get a dynamic binary like so: [bandix@scylla amanda-2.4.2p2]% ldd client-src/amandad libm.so.6 = /lib/i686/libm.so.6 (0x4002b000) libreadline.so.4 = /usr/lib/libreadline.so.4 (0x4004e000) libtermcap.so.2 = /lib/libtermcap.so.2 (0x40074000) libnsl.so.1 = /lib/libnsl.so.1 (0x40078000) libc.so.6 = /lib/i686/libc.so.6 (0x4008e000) /lib/ld-linux.so.2 = /lib/ld-linux.so.2 (0x4000) The only substantial difference between GNUtar and amanda's build processes I can discern is that amanda is using libtool. Any recommendations on how to get amanda convinced to compile everything statically? Brandon D. Valentine -- Iam mens praetrepidans avet vagari. - G. Valerius Catullus, Carmina, XLVI
tapetype recommendations
Well, today for the first time I hit EOT. I'm using a Sony SDX-300C
AIT-1 drive. Here's the tapetype I've been using and the error messages
as produced by amreport:
define tapetype AIT1-35 {
comment AIT-1 with 230m tapes (35/70 GB)
length 33000 mbytes
filemark 500 kbytes
speed 3000 mbytes
}
--- amreport email ---
These dumps were to tape Daily06.
*** A TAPE ERROR OCCURRED: [[writing file: No space left on device]].
Some dumps may have been left in the holding disk.
Run amflush to flush them to tape.
The next tape Amanda expects to use is: Daily07.
snip
STATISTICS:
Total Full Daily
Estimate Time (hrs:min)0:03
Run Time (hrs:min) 4:16
Dump Time (hrs:min)7:05 5:13 1:52
Output Size (meg) 46003.234804.911198.2
Original Size (meg) 76399.859044.717355.1
Avg Compressed Size (%)60.2 58.9 64.5 (level:#disks
...)
Filesystems Dumped 44 15 29 (1:28 4:1)
Avg Dump Rate (k/s) 1846.3 1896.9 1705.0
Tape Time (hrs:min)3:00 2:55 0:05
Tape Size (meg) 28873.028244.6 628.5
Tape Used (%) 87.6 85.61.9 (level:#disks
...)
Filesystems Taped41 14 27 (1:27)
Avg Tp Write Rate (k/s) 2739.3 2751.7 2278.3
snip
taper: tape Daily06 kb 34136032 fm 42 writing file: No space left on
device
driver: going into degraded mode because of tape error.
--- end amreport email ---
Any recommendations on how to change my tapetype definition to avoid
this problem? And yes, I've got hardware compression disabled.
Brandon D. Valentine
--
Iam mens praetrepidans avet vagari.
- G. Valerius Catullus, Carmina, XLVI
RE: tapetype recommendations
On Wed, 5 Dec 2001, Anthony Valentine wrote: Your tapes hold 35GB native. Correct. You are backing up: 76399.8 MB (76GB) Compressed, that is: 46003.2 MB (46GB) Because you use software compression, you have to go by the native capacity of your tapes. So even compressed, you data won't fit on one tape. My quarrel seems to be with the fact that amanda is attempting to backup more data than will fit on a single tape. This might not be a tapetype problem, but it seems the logical place to start twiddling. It may be that planner is unable to get good estimates for some of this data so what it thought would compress to 35GB only became 46GB post dump. The only options that I can see are to try to get better compression by using 'gzip --best' (if Amanda can support adding switches to gzip) or getting larger tapes. This is unacceptable as I have found that gzip --best, which amanda does support, causes a nearly exponential increase in the time it takes amdump to run. A considerable fraction of my clients are r5k/r10k/r12k MIPS machines and none of them are terribly fast at gzip as it is. Few of the GNU tools are particularly well optimized for the SGI MIPS platform. Larger tapes are also not available for AIT-1. Setting the length option in the tapetype clause to more than your tape can hold probably won't help you any. Correct me if I'm wrong but what I have set the length option to is well _below_ what the tape will hold. With 35GB capacity tapes I have 35000mb or 35840mb depending upon whether Sony uses powers of 2 (the computer science definition) or powers of 10 (the marketing definition) when defining their tape size. I'll go with the former to be safe. I've set length in tapetype to 33000mb, a good 2GB below what the tape will theoretically hold. Worst case scenario is that Amanda uses powers of 2 to define tape size and Sony uses powers of 10. If this is the case then 33000*1024*1024/1000/1000 = 34603 so I should still be safe. Brandon D. Valentine -- Iam mens praetrepidans avet vagari. - G. Valerius Catullus, Carmina, XLVI
Re: [Amanda-users] Tried this before....
On Thu, 6 Dec 2001, Jason Thomas wrote: far enough, its just that usually when people send unsubscribes to the actual list they are idiot. I now see this is not the case. so now you should send an email to: It's worth commenting that several months ago I had the same issue unsubscribing an adress from amanda-users. While I did not inquire to the list about it, I did send mail to the list admin who was very responsive in getting the situation taken care of. I am not sure why amanda's majordomo configuration has trouble with unsubscribes. I got my email through to the list admin via [EMAIL PROTECTED], a guess at a pretty standard majordomo alias which forwards to a list admin. Brandon D. Valentine -- Iam mens praetrepidans avet vagari. - G. Valerius Catullus, Carmina, XLVI
Re: Redhat 7.1 troubles
On Tue, 23 Oct 2001, Jennifer Peterson wrote: after about 20 minutes or so, I get this in my syslog: Oct 22 15:24:11 my_computer xinetd[29480]: amanda service was deactivated because of looping Oct 22 15:24:41 my_computer amandad[30051]: error receiving message: timeout The amandad error log on my_computer reads basically the same thing: amandad: dgram_recv: timeout after 30 seconds amandad: error receiving message: timeout Have you tried looking at the latest debug output in /tmp/amanda after running amcheck? I ran into the deactivated because of looping error on a few clients which were still GLIB2.1 based machines where my NFS automounted amanda binaries were compiled against GLIBC2.2. Find out what errors the amandad binary is spewing into /tmp/amanda/*.debug -- Never put off until tomorrow what you can do today. There might be a law against it by that time. -- /usr/games/fortune, 07/30/2001 Brandon D. Valentine bandix at looksharp.net
Statically linked binaries?
Greetings, I'm running into trouble when trying to get Amanda binaries statically linked. I'm trying to compile a set of statically linked linux binaries that I can place on an NFS mount where linux systems will run them without fighting over libc revisions, etc. When I run ./configure with the --enable-static=yes option I still end up with dynamically linked binaries, as revealed by running ldd after a make install. Setting the CFLAGS environment variable to contain -static doesn't help either. How can I get Amanda to produce statically linked binaries for me? TIA, -- Never put off until tomorrow what you can do today. There might be a law against it by that time. -- /usr/games/fortune, 07/30/2001 Brandon D. Valentine bandix at looksharp.net
