Re: amlabel ftape
hello johannes, tx for the hint, but it won't work. still the same probs. if you send me the configuration, i will appreciate it. greets stefan
About reporter.c output formatting, was Re: Over large level 1 dumps
I was out of pocket for 2 weeks, doing the Grand Old Man scene at a sister station, and when I got back I grabbed 2.4.3b3-20020422 and built it without doing my usual edit job on reporter.c to make its output readable as it seemed reporter.c had been heavily re-written, according to the ChangeLog, by John R. Jackson about 2002-04-07, after I had caught a plane to Grand Junction. The new reporter.c formatting is MUCH better, but it still has a wibble at the 3rd column from the right end of a line when doing the 'host directory' etc lines in the mailed report. It appears that this 3rd from the end column is miss-calculated for its size by 1 char short, pulling the 2nd from the end in by 1 char if the 3rd column contains a short value, and has a permanent 2 char spacing between the 3rd and 2nd from the end columns. So in the report, this 3rd column back is subjected to a variable width, and the last 2 columns then are staggered back and forth. This is a minor quibble really as this format is much easier to read, its more of a heads up to John. I walked around in it a bit, but didn't quite get my head into the newer code. I'd need more sleep before I tackled that, or a younger brain, this ones getting tired. :) -- Cheers, Gene AMD K6-III@500mhz 320M Athlon1600XP@1400mhz 512M 98.85+% setiathome rank, not too shabby for a hillbilly
build and install for different machines.
Hi, I have just installed amanda and run it up and I have to say I'm really impressed. But you all know it's good, so I won't go on. One thing I have noticed, though, does not appear to be covered in the docs. Our shop runs a small number of development systems with compilers, and a larger number of machines on which I will want to install the client only. Furthermore none of the development machines are the backup server. What I want to do is build the client and server pieces (for suns, in case you're wondering) on one machine, and distribute them to where they have to go. I don't need much customisation for each machine, just something I can plonk down and install. There does not appear to be a neat way to build a gzip, tarball or solaris package of the server or the client for installation somewhere else. I don't particularly want to install a development environment on the backup server, and in any case this wouldn't solve the client part. A second part of this project will be to compile a linux client-only piece and do the same thing for the linux boxes, none of which is a tape host. So I'd need info on how to do that too. All the systems will be pointing at the one sun backup server. Doubtless this has already been sorted out by some clever person but I can't locate the information. Any pointers would be greatly appreciated. Chuck
Over large level 1 dumps
I asked about this the other day and it transpired that I didn't have record yes in my dumptype, so that explained that, or so I thought. However, I now do have record yes in my global dumptype but still level 1 dumps are of the full filesystems. amadmin TIZ find no gives me this list of backups Scanning /data/amandahold... date host disk lv tape or file file status 2002-04-16 moca / 0 TIZdaily01 3 OK 2002-04-18 moca / 1 TIZdaily02 3 OK 2002-04-20 moca / 1 TIZdaily03 3 OK 2002-04-23 moca / 1 TIZdaily04 3 OK 2002-04-24 moca / 1 TIZdaily05 3 OK 2002-04-16 moca /boot 0 TIZdaily01 1 OK 2002-04-18 moca /boot 1 TIZdaily02 4 OK 2002-04-20 moca /boot 1 TIZdaily03 4 OK 2002-04-23 moca /boot 0 TIZdaily04 4 OK 2002-04-24 moca /boot 0 TIZdaily05 4 OK 2002-04-16 pumori / 0 TIZdaily01 4 OK 2002-04-18 pumori / 0 TIZdaily02 1 OK 2002-04-20 pumori / 1 TIZdaily03 1 OK 2002-04-23 pumori / 1 TIZdaily04 1 OK 2002-04-24 pumori / 1 TIZdaily05 1 OK 2002-04-16 pumori /boot 0 TIZdaily01 2 OK 2002-04-18 pumori /boot 1 TIZdaily02 2 OK 2002-04-20 pumori /boot 0 TIZdaily03 2 OK 2002-04-23 pumori /boot 0 TIZdaily04 2 OK 2002-04-24 pumori /boot 1 TIZdaily05 2 OK moca:/etc/amandates is / 1 1019600282 /boot 0 1019600788 and pumori:/etc/amandates is / 1 1019604259 /boot 1 1019605208 Another thought just occured to me - as I wasn't recording, why did amanda ever do level 1 backups ? My tape is big enough to hold all of the filesystems (currently in disklist - I do want to add more) at level 0 so it can't have been a size issue. And is it possible for me to test this by doing more than one backup run in a day - isn't there an issue with amanda only recording certain things by date and not by time. If I had 'index no' in my dumptype would I get around that ? Regards, Niall O Broin
Re: What to do when tape is full?
On 23/04 2002 17:38 Frank Smith wrote: --On Tuesday, April 23, 2002 15:26:33 +0200 Toralf Lund [EMAIL PROTECTED] wrote: On 23/04 2002 14:52 Joshua Baker-LePain wrote: Yes, maybe a good decision. (Although the purchase of new tapes and the management of all of them do add up to a non-negligible cost.) While the cost of sufficient tapes is certainly 'non-negligible', and can actually be much greater than the cost of the library you put them, company management has to weigh the cost of tapes versus the cost of losing data. Sooner or later you will need to restore something. How much is it worth to be able to do that? There is much data that can be (somewhat) easily recreated, and other data that is irreplaceable or would take so long to recreate that you would be out of business before it could be recreated. If you don't have enough tapes to be comfortable in the ability to restore your data (and just because you have one copy doesn't mean you will be able to restore it), then get more tapes. If your management doesn't see the need, write up what you need and why, and be sure to save the response if they disagree, since your head is on the line if critical data can't be restored. Many good points, there. These are exactly the considerations my backup configuration is based on, really. I suspect that a lot of people back up their data because conventional wisdom is that they should do it, whereas the backup strategy really ought to be based on an assessment of the cost of backup and recovery vs the cost of recreating data. - Toralf
Re: amcheck
On Tue, 23 Apr 2002 at 3:17pm, Michael Hall wrote Amanda Backup Client Hosts Check ERROR: bzncs03: [could not access sda1 (sda1): No such file or directory] Client check: 2 hosts checked in 0.046 seconds, 1 problem found (brought to you by Amanda 2.4.2p2) You need to apply the advfs.patch from http://www.amanda.org/patches.html. Apply that, recompile, reinstall, and all will be well with your world. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Over large level 1 dumps
On Wed, 24 Apr 2002 at 12:02pm, Niall O Broin wrote I asked about this the other day and it transpired that I didn't have record yes in my dumptype, so that explained that, or so I thought. However, I now do have record yes in my global dumptype but still level 1 dumps are of the full filesystems. amadmin TIZ find no gives me this list of backups But if you've *just* turned 'record yes' on, then it should take a dumpcycle for things to work themselves out (I think). -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Couple of Log File Questions
On Wed, 24 Apr 2002 at 1:12pm, GIC MLs wrote A couple of questions from my log file: planner:RDEV_PREFIX=/dev/r DUMP=/sbin/dump planner:RESTORE=/sbin/restore GNUTAR=/usr/bin/tar Should I be using gtar instead? Yes -- you *must* use GNU tar. But, depending upon your OS, /usr/bin/tar, may be GNU. Try '/usr/bin/tar --version'. How do I change this setting? Install GNUtar (if necessary), then recompile amanda specifying the correct location. planner:DEFAULT_CONFIG=DailySet1 My amanda.conf has labelstr ^daily[0-9][0-9]$* - where is this DailySet1 coming from? planner:DEFAULT_TAPE_DEVICE=/dev/null HAVE_MMAP HAVE_SYSVSHM My amanda.conf has tapedev /dev/nrsa0 ... what gives? These are defaults compiled into the various amanda bits. Look at the output of './configure --help' in the amanda source directory to see where the defaults come from. got result for host client.agic.ne.jp disk /usr: 0 - -1K, -1 - -1K, -1 - -1K Is this bad? Yes. And the failure that I've been getting: getting estimates took 0.183 secs FAILED QUEUE: 0: client.agic.ne.jp /usr DONE QUEUE: empty ANALYZING ESTIMATES... planner: FAILED client.agic.ne.jp /usr 0 [disk /usr offline on client.agic.ne.jp?] It's probably a permissions issue. What user/group are you running amanda as? Are you using inetd or xinetd? Did you do the install as root? What OS? -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Estimates - 7 hour 50mins
On 23 Apr 2002, at 9:37, Joshua Baker-LePain wrote: I found one posting dealing with how to use it -- it contains a modified sendsize.c that it seems like you plop into place and recompile:\ http://marc.theaimsgroup.com/?l=amanda-usersm=9253924842w=2 I think this sendsize.c was maybe for amanda2.4.2p2 as I got errors when compiling it with amanda2.4.3b3 which is the version I use. This leaves me with the following questions: 1. Is there a version of sendsize.c that is compatibile with amanda 2.4.3b3? 2. Is compiling the source with an updated sensize.c the only method of using calcsize? The other suggestion I ran across was to use ufsdump rather than tar. Is there a reason you need to use tar? - David Flood Systems Administrator [EMAIL PROTECTED] Tel: +44 (0)1224 262721 The Robert Gordon University School of Computing St. Andrews Street Aberdeen -
Re: Estimates - 7 hour 50mins
On Wed, 24 Apr 2002 at 2:03pm, David Flood wrote On 23 Apr 2002, at 9:37, Joshua Baker-LePain wrote: I found one posting dealing with how to use it -- it contains a modified sendsize.c that it seems like you plop into place and recompile:\ http://marc.theaimsgroup.com/?l=amanda-usersm=9253924842w=2 I think this sendsize.c was maybe for amanda2.4.2p2 as I got errors when compiling it with amanda2.4.3b3 which is the version I use. This leaves me with the following questions: 1. Is there a version of sendsize.c that is compatibile with amanda 2.4.3b3? I don't know that there's a modified sendsize.c for 2.4.3. *However*, is there a reason you need 2.4.3b3? You realize 'b' is for beta, right? 2. Is compiling the source with an updated sensize.c the only method of using calcsize? That I know of -- yes. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Estimates - 7 hour 50mins
On 24 Apr 2002, at 9:10, Joshua Baker-LePain wrote: 1. Is there a version of sendsize.c that is compatibile with amanda 2.4.3b3? I don't know that there's a modified sendsize.c for 2.4.3. *However*, is there a reason you need 2.4.3b3? You realize 'b' is for beta, right? Yes I know b is for beta but I was upgrading my other tapeserver to 3b3 for use of the extended samba exclusion lists introduced in this version. So I upgraded both my servers at the same time but I suppose since the one causing the problems does not use any excludes especially samba I could put it back to version 2p2. David Flood Systems Administrator
Re: amcheck
That solved the problem Thank you for your help. -mike Joshua Baker-LePain wrote: On Tue, 23 Apr 2002 at 3:17pm, Michael Hall wrote Amanda Backup Client Hosts Check ERROR: bzncs03: [could not access sda1 (sda1): No such file or directory] Client check: 2 hosts checked in 0.046 seconds, 1 problem found (brought to you by Amanda 2.4.2p2) You need to apply the advfs.patch from http://www.amanda.org/patches.html. Apply that, recompile, reinstall, and all will be well with your world. -- Michael Hall, [EMAIL PROTECTED] Systems Administrator Zoot Enterprises Inc. www.zootweb.com
Re: Over large level 1 dumps
On Wed, Apr 24, 2002 at 12:02:54PM +0100, Niall O Broin wrote: I asked about this the other day and it transpired that I didn't have record yes in my dumptype, so that explained that, or so I thought. However, I now do have record yes in my global dumptype but still level 1 dumps are of the full filesystems. Correct. Last night, all your filesystems thought they'd never had a level 0 dump (since none had been recorded), so level 1 included everything. As each filesystem does (and records) a level 0 dump, it will start doing incremental level 1s. After a full dumpcycle, things should be behaving as you expect. Another thought just occured to me - as I wasn't recording, why did amanda ever do level 1 backups ? Because it wasn't time for a level 0. Full (lvl 0) vs. incrmental (lvl. 1/2) is controlled by the dumpcycle and ensuring that level 0s are done often enough to always have one within the last dumpcycle. If a lvl 0 isn't needed, a lvl 1 or 2 will be done instead regardless of how big the dumps are estimated at.
sendsize.c for calcsize
On 24 Apr 2002, at 14:20, David Flood wrote: On 24 Apr 2002, at 9:10, Joshua Baker-LePain wrote: 1. Is there a version of sendsize.c that is compatibile with amanda 2.4.3b3? I don't know that there's a modified sendsize.c for 2.4.3. *However*, is there a reason you need 2.4.3b3? You realize 'b' is for beta, right? Yes I knew b was for beta, I am using 3b3 because of it's exclusion file capability with samba. This is on the other tape server but I upgraded them both at the same time. There is no reason why the tapeserver causing the problems has to be 3b3. With this in mind I tried to recompile 2p2 with the sensize.c that makes use of calcsize but it still gave me errors when running make. I've attached the sendsize.c and the list of errors I got incase anyone is feeling bored. I've never programmed in C nor have enough experience with amanda to check this file. Failing this does anyone have a sendsize.c that works? David Flood Systems Administrator
Re: Amanda + solaris
Solaris 8 does. Thomas Hepper wrote: [EMAIL PROTECTED]"> Hi,On Tue, Apr 23, 2002 at 03:35:34PM +0200, Ryszard Kluza wrote: HelloI have Sun E250 with Solaris 7Tape library Adic FastStore22And amandaHow I have to configure amanda and solaris to work together,What kind of driver I need to instal on Solaris I'm not sure if solaris have already the sgen driver. If no check the docs directory and there the file TAPE.CHANGERS on what youneed for solaris without the sgen driver. In this doc you will findalso some notes on how to configure the sgen driver.Hope this helps for the first steps. Thomas -- _-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_- Don Potter -- [EMAIL PROTECTED] Center for Human Genetics Duke University -_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_ Today's subliminal thought is:
gnutar options
I see this in the amanda.conf file: # Note that the `full pathname' of a file within its # filesystem starts with `./', because of the way amanda runs # gnutar: `tar -C $mountpoint -cf - --lots-of-options .' I was wondering how I can find out what the lots-of-options are? Thanks, -- Terri Eads Systems Administrator [EMAIL PROTECTED] Research Applications Program (303) 497-8425 National Center for Atmospheric Research
Re: gnutar options
On Wed, 24 Apr 2002 at 11:40am, Terri Eads wrote I see this in the amanda.conf file: # Note that the `full pathname' of a file within its # filesystem starts with `./', because of the way amanda runs # gnutar: `tar -C $mountpoint -cf - --lots-of-options .' I was wondering how I can find out what the lots-of-options are? After an amdump run, look in /tmp/amanda/runtar*debug -- you'll see the exact command line. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Estimates - 7 hour 50mins
Various people I found one posting dealing with how to use it -- it contains a modified sendsize.c that it seems like you plop into place and recompile:\ http://marc.theaimsgroup.com/?l=amanda-usersm=9253924842w=2 I think this sendsize.c was maybe for amanda2.4.2p2 as I got errors when compiling it with amanda2.4.3b3 which is the version I use. This leaves me with the following questions: 1. Is there a version of sendsize.c that is compatibile with amanda 2.4.3b3? I don't know that there's a modified sendsize.c for 2.4.3. *However*, is there a reason you need 2.4.3b3? You realize 'b' is for beta, right? If this is the modifcation I made long agao, then no, there is no mod for 2.4.3 2. Is compiling the source with an updated sensize.c the only method of using calcsize? That I know of -- yes. I was once, long ago, undertaking the effort of sticking this in the proper amanda tree. But doing this proper was (and still is) a lot more work then I have time for. There is also an issue where the sendsize/calcsize combo gives incorrect results, and although in my particular case (500G of disks to be backup on a 100G LTO tape the errors are small enough to be non-significant I have seen posts here in this forum where the error was big enough to give serious incorrect estimats. Also, this particular hack does not support exclude patterns The problem lies in the way incremental dumpsizes are calculated, and either lists incrementals as having size 0 (32K) while they are in fact much bigger or lists incrementals as being almost level 0 size while they are not. Given that Im dumping 41 disks/tar-slices in a one week cycle I have on average 70% of my backups as level 0 dump (8 or 9 disks/slices) and the rest incrementals means that the total error between the calcsize estimate and the real dumpsize is less than the amount of data being generated between the start of the dump and the end. (amanda runs for 8+ hours, in that period my users can eaily generate over 10G of data, the joys of working with seismic processors ;) ) So for me those errors are not an issue. Having said all that, I can give you a sendsize binary (2.4.2p1 I think) for SOlaris 2.6 (and higher), Irix 6.5 and Linux (2.4 kernel series). Or if you can give me the compile log I can see what is causing your problem. (sorry for not picking upt his thread earlier, I don;t even have enough time to follow the amanda mailinglist these days ..) Kind regards, -- Gerhard den Hollander Phone :+31-10.280.1515 Global IT Support manager Direct:+31-10.280.1539 Jason Geosystems BV Fax :+31-10.280.1511 (When calling please note: we are in GMT+1) [EMAIL PROTECTED] POBox 1573 visit us at http://www.jasongeo.com 3000 BN Rotterdam JASON...#1 in Reservoir CharacterizationThe Netherlands This e-mail and any attachment is/are intended solely for the named addressee(s) and may contain information that is confidential and privileged. If you are not the intended recipient, we request that you do not disseminate, forward, distribute or copy this e-mail message. If you have received this e-mail message in error, please notify us immediately by telephone and destroy the original message.
Re: gnutar options
* Joshua Baker-LePain [EMAIL PROTECTED] (Wed, Apr 24, 2002 at 02:06:51PM -0400) On Wed, 24 Apr 2002 at 11:40am, Terri Eads wrote I see this in the amanda.conf file: # Note that the `full pathname' of a file within its # filesystem starts with `./', because of the way amanda runs # gnutar: `tar -C $mountpoint -cf - --lots-of-options .' I was wondering how I can find out what the lots-of-options are? After an amdump run, look in /tmp/amanda/runtar*debug -- you'll see the exact command line. Which in my case is: (reformatted for readability XXX is the name of the dir you listed in disklist the name after listed incremental hyperionsdb2_1.new is hostname (hyperion) disk (sdb2) runtar: debug 1 pid 10020 ruid 0 euid 0 start time Tue Apr 16 20:05:06 2002 gtar: version 2.4.2p1 running: /bin/gtar: gtar --create --directory XXX \ --listed-incremental \ /usr/local/var/amanda/gnutar-lists/hyperionsdb2_1.new \ --sparse --one-file-system --ignore-failed-read \ --totals --file - . Gerhard, @jasongeo.com == The Acoustic Motorbiker == -- __O Some say the end is near. =`\, Some say we'll see armageddon soon (=)/(=) I certainly hope we will I could use a vacation
Re: gnutar options
Gerhard den Hollander wrote: * Joshua Baker-LePain [EMAIL PROTECTED] (Wed, Apr 24, 2002 at 02:06:51PM -0400) On Wed, 24 Apr 2002 at 11:40am, Terri Eads wrote I see this in the amanda.conf file: # Note that the `full pathname' of a file within its # filesystem starts with `./', because of the way amanda runs # gnutar: `tar -C $mountpoint -cf - --lots-of-options .' I was wondering how I can find out what the lots-of-options are? After an amdump run, look in /tmp/amanda/runtar*debug -- you'll see the exact command line. Which in my case is: (reformatted for readability XXX is the name of the dir you listed in disklist the name after listed incremental hyperionsdb2_1.new is hostname (hyperion) disk (sdb2) runtar: debug 1 pid 10020 ruid 0 euid 0 start time Tue Apr 16 20:05:06 2002 gtar: version 2.4.2p1 running: /bin/gtar: gtar --create --directory XXX \ --listed-incremental \ /usr/local/var/amanda/gnutar-lists/hyperionsdb2_1.new \ --sparse --one-file-system --ignore-failed-read \ --totals --file - . Gerhard, @jasongeo.com == The Acoustic Motorbiker == -- __OSome say the end is near. =`\,Some say we'll see armageddon soon (=)/(=) I certainly hope we will I could use a vacation I don't have a /tmp/amanda/runtar*debug file. Do I have to already be running tar to generate this file? -- Terri Eads Systems Administrator [EMAIL PROTECTED] Research Applications Program (303) 497-8425 National Center for Atmospheric Research
Re: gnutar options
On Wed, 24 Apr 2002 at 1:44pm, Terri Eads wrote I don't have a /tmp/amanda/runtar*debug file. Do I have to already be running tar to generate this file? Yes. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
not really amanda problem: tar, Cannot stat
I decided to start using tar for a particular machine but received the following for one of the file systems (the rest of the machine backed up without problems): FAILURE AND STRANGE DUMP SUMMARY: utnetdb.gw /usr lev 0 FAILED [/usr/local/bin/tar returned 2] /-- utnetdb.gw /usr lev 0 FAILED [/usr/local/bin/tar returned 2] sendbackup: start [utnetdb.gw.utexas.edu:/usr level 0] sendbackup: info BACKUP=/usr/local/bin/tar sendbackup: info RECOVER_CMD=/usr/local/bin/tar -f... - sendbackup: info end ? gtar: ./X11R6/share/locale/fr: Cannot stat: Input/output error ? gtar: ./home/uttelnetuser/.login: Cannot stat: Input/output error ? gtar: ./home/uttelnetuser/.login_conf: Cannot stat: Input/output error ? gtar: ./home/uttelnetuser/.mailrc: Cannot stat: Input/output error ? gtar: ./home/uttelnetuser/.profile: Cannot stat: Input/output error ? gtar: ./home/uttelnetuser/.shrc: Cannot stat: Input/output error ? gtar: ./home/uttelnetuser/.mail_aliases: Cannot stat: Input/output error ? gtar: ./home/uttelnetuser/.rhosts: Cannot stat: Input/output error ? gtar: ./local/include/tcl8.3/unix: Cannot stat: Input/output error ? gtar: ./local/lib/expect5.32/libexpect532.a: Cannot stat: Input/output error ? gtar: ./local/lib/expect5.32/pkgIndex.tcl: Cannot stat: Input/output error ... ... ... | Total bytes written: 1635614720 (1.5GB, 235kB/s) ? gtar: Error exit delayed from previous errors sendbackup: error [/usr/local/bin/tar returned 2] \ Any thoughts on what this means? Tried searching through the archives and found one or two questions on this but no answers. Oscar
Elaboration on tape changer behavior
Can those with experience with changers elucidate some details for me? What is the AMANDA/amtape policy when rolling over from the last tape to the first tape in my Backup Set? My config specifies the following: dumpcycle 2 weeks runspercycle 5 tapecycle 10 tapes In my changer the slot numbering is zero-based. So I have tapes in slots 0 thru 9. Since I do not have a working changer script between amtape and my tape library manager sw layer, I am manually running robot load commands (etc...). I have three general questions: 1) Will amtape manage the rollover from tape slot 9 to slot 0, and, in general, do the right thing? Is there any manual intervention on tape slot rollover? For example. must the admin tell AMANDA that it is OK to write over a tape (tape 00)? 2) Is there any important information omitted from the document, AMANDA Tape Changer Support? Can it generally be trusted to specify the correct interfaces? I have found that some of the tape changer implementations in the AMANDA distribution do not correctly follow the API specifications in that document. 3) Do people with single tape robots use multiple tape magazines; swapping them out after one tapecycle... as in Magazine A then Magazine B, then back to A? Or do some people just keep running tapes through the magazine, ad infinitum? Or do most people leave the same magazine installed until the tapes wear out? I presume this last question (3) is largely dependent upon one's goals... So if anybody cares to elaborate on changer methodologies used to achieve particular goals, then I would be grateful. Other will surely benefit as this is not an area that is discussed in much depth. Thanks for your time... J
Re: build and install for different machines.
On Wed, Apr 24, 2002 at 06:31:10PM +1000, [EMAIL PROTECTED] wrote: Our shop runs a small number of development systems with compilers, and a larger number of machines on which I will want to install the client only. Furthermore none of the development machines are the backup server. What I want to do is build the client and server pieces (for suns, in case you're wondering) on one machine, and distribute them to where they have to go. I don't need much customisation for each machine, just something I can plonk down and install. There does not appear to be a neat way to build a gzip, tarball or solaris package of the server or the client for installation somewhere else. I don't particularly want to install a development environment on the backup server, and in any case this wouldn't solve the client part. It's really just a matter of making sure that all of the backup tools amanda will be calling on any of the clients (ufsdump, gtar, vxdump, etc.) are present in the same location on the build machine and specifying the configure options (--with-tape-server, --with-tape-device, etc.) to reflect the environment amanda will be running in. Build and install as normal on your build machine, and tar or package up the results. -- Mike Delaney [EMAIL PROTECTED] ...Microsoft follows standards. In much the same manner that fish follow migrating caribou. Now I have this image in my mind of a fish embracing and extending a caribou. -- Paul Tomblin and Christian Bauernfeind in the SDM
Re: Elaboration on tape changer behavior
--On Wednesday, April 24, 2002 13:23:08 -0700 John Koenig [EMAIL PROTECTED] wrote: Can those with experience with changers elucidate some details for me? What is the AMANDA/amtape policy when rolling over from the last tape to the first tape in my Backup Set? My config specifies the following: dumpcycle 2 weeks runspercycle 5 tapecycle 10 tapes In my changer the slot numbering is zero-based. So I have tapes in slots 0 thru 9. Since I do not have a working changer script between amtape and my tape library manager sw layer, I am manually running robot load commands (etc...). I have three general questions: 1) Will amtape manage the rollover from tape slot 9 to slot 0, and, in general, do the right thing? Is there any manual intervention on tape slot rollover? For example. must the admin tell AMANDA that it is OK to write over a tape (tape 00)? Yes, it knows to roll over from the last tape to the first, but see #3 below. 2) Is there any important information omitted from the document, AMANDA Tape Changer Support? Can it generally be trusted to specify the correct interfaces? I have found that some of the tape changer implementations in the AMANDA distribution do not correctly follow the API specifications in that document. I'll pass on this one. 3) Do people with single tape robots use multiple tape magazines; swapping them out after one tapecycle... as in Magazine A then Magazine B, then back to A? Or do some people just keep running tapes through the magazine, ad infinitum? Or do most people leave the same magazine installed until the tapes wear out? I presume this last question (3) is largely dependent upon one's goals... So if anybody cares to elaborate on changer methodologies used to achieve particular goals, then I would be grateful. Other will surely benefit as this is not an area that is discussed in much depth. You need at least a couple more tapes, since with your current setup if you ever have a tape error or an extra-large backup and need to amflush, it would need to overwrite an active tape. The other answer is, as you suspect, 'it depends'. If you do offsite rotations of your tapes, it is easier to do it a magazine at a time. If your robot holds several magazines you might want to cycle through one set for onsite backups and have a different Amanda config do offsite backups on another magazine that you replace each time. Even if you don't do offsite, swapping tapes a magazine at a time is much safer than swapping tapes. The tapes will last longer in a magazine (less likely to get banged around, dropped, dirtied, etc.), and you have fewer chances to get the tapes out of order. Magazines also stack better and are easier to store. I'm assuming your magazines can be swapped easily, if not then it might not be as favorable to swap them. Unfortunately, some vendors see magazines as an income opportunity and charge way too much for them. Also, do you need archival backups (to retrieve a file that was only around last June), or only the ability to restore what was there in the last few days or weeks. With your config, if a backup fails for some reason, and a file you need wasn't on the system in the last week, you may not have it (as you are just now about to overwrite your first tape, and whatever fulls it contains cannot be recreated with the following 4 incrementals so all you can recover from those filesytems of that week is the changed files. If the full of those filesystems can't be read from the one this week (due to a bad tape or whatever) you are SOL. I would recommend making sure you always have at least two full backups of everything, even if there are a couple of errors during your cycle. Good luck, Frank -- Frank Smith[EMAIL PROTECTED] Systems Administrator Voice: 512-374-4673 Hoover's Online Fax: 512-374-4501
Re: Over large level 1 dumps
On Wednesday, April 24, 2002, at 08:26 AM, Dave Sherohman wrote: On Wed, Apr 24, 2002 at 12:02:54PM +0100, Niall O Broin wrote: I asked about this the other day and it transpired that I didn't have record yes in my dumptype, so that explained that, or so I thought. However, I now do have record yes in my global dumptype but still level 1 dumps are of the full filesystems. Correct. Last night, all your filesystems thought they'd never had a level 0 dump (since none had been recorded), so level 1 included everything. As each filesystem does (and records) a level 0 dump, it will start doing incremental level 1s. After a full dumpcycle, things should be behaving as you expect. But the man page says the default value for record is yes. What is up? EZ
[Fwd: amcheck output]
---BeginMessage--- HI, I am running amcheck on my backup called Daily and I am getting the following output: Amanda Tape Server Host Check - ERROR: tapelist dir /usr/local/share/amanda/Daily: not writable Holding disk /amanda: 404144 KB disk space available, that's plenty (expecting a new tape) NOTE: skipping tape-writable test NOTE: info dir /usr/local/share/amanda/Daily/curinfo: does not exist NOTE: it will be created on the next run NOTE: index dir /usr/local/share/amanda/Daily/index: does not exist Server check took 9.687 seconds Amanda Backup Client Hosts Check ERROR: stud.hailix.com: [can not read/write /usr/local/var/amanda/gnutar-lists/.: No such file or directory] Client check: 1 host checked in 0.028 seconds, 1 problem found I am using the root-tar dump type. I just installed the gnu tar and reconfigured amanda. I put no where in the amanda.conf file any such directory /usr/local/var/amanda/gnutar-lists. Amanda and associated progarms are ing /usr/local/etc/amanda. Any thoughts? TIA Trevor ---End Message---
