HP LTO Ultrium changer plus timeout problems
Hi, you couldn't by any chance give your conf file could you? We tried for months trying to get the tape changer to work but it never happened. We just manually change them each day, waiting for the day that we purchase an expensive backup tool. Also, is there a way to stop it from estimating? We use full backups (130Gb) every night and the thing spends hours running it's compression estimations (even though it is turned off), the result of which is that it sometimes times out and doesn't back any thing up from large file systems (50K+ directories). The simplest method would be to tell it to use "du -sk" but so far, I haven't had the time. Any help would be most welcome. cheers . --- In [EMAIL PROTECTED], Gerhard den Hollander wrote: >* Lylace Garcia-Blake (Wed, Feb 07, 2001 at 08:28:37AM -0800) > > Happy Wednesday, > > > Does someone have a tapetype entry for the HP LTO Ultrium device that > > they would be willing to post? > >no tapetype (at least non official) >but the following works for me > > >define tapetype LTO { >comment "LTO ultrium" >length 150 gbytes # conservative estimate >filemark 1 byte # should work given above >speed 30 mbytes # even more, but this isn't used in amanda ># yet >lbl-templ "/volume/amanda/share/lto.ps" >} > > >Notes: >if you use compression on amanda (iso using drive compression) set this to >100G. > >With our data the LTO drive does a bit better than 1:1.5 compression >(I assume 1.6 but I'd rather err on the side of caution) > > > >Gerhard, <@jasongeo.com> == The Acoustic Motorbiker == >-- >__O If your watch is wound, wound to run, it will >=`\<, If your time is due, due to come, it will >(=)/(=) Living this life, is like trying to learn latin >in a chines _ MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx
amanda in aix
Hello: I'm trying to compile amanda 2.4.2p2 in a 4.3.3 aix system. i installed gtar,gawk, gsed, perl and readline. i run the configure script, but when i do make i receive the next error: make[1]: Entering directory `/home/software/amanda-2.4.2p2/client-src' /usr/bin/sh ../libtool --mode=link gcc -g -O2-o amandad amandad.o ../common-src/libamanda.la libamclient.la ../common-src/libamanda.la -lm -lreadline -lcurses -lnsl -lintl gcc -g -O2 -o amandad amandad.o ../common-src/.libs/libamanda.a -lm -lreadline -lcurses -lnsl -lintl .libs/libamclient.a -lm -lreadline -lcurses -lnsl -lintl ../common-src/.libs/libamanda.a -lm -lreadline -lcurses -lnsl -lintl -lm -lreadline -lcurses -lnsl -lintl ld: 0711-317 ERROR: Undefined symbol: .__main ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information. collect2: ld returned 8 exit status make[1]: *** [amandad] Error 1 make[1]: Leaving directory `/home/software/amanda-2.4.2p2/client-src' make: *** [all-recursive] Error 1
FAILURE AND STRANGE DUMP SUMMARY:
Hi Helpers, as you can see, I am currently backing up just one client using gnutar. Now, having one tapecycle run, suddenly 3 of 14 mountpoints on the same machine failed for reasons I don't get. The mail from today, report from the night 25.-26. ---cut-on--- ... machine /some/mountpoint_1 lev 1 FAILED [could not connect to machine] machine /some/mountpoint_2 lev 1 FAILED [could not connect to machine] machine /some/mountpoint_3 lev 1 FAILED [could not connect to machine] STATISTICS: ... DUMP SUMMARY: DUMPER STATSTAPER STATS HOSTNAME DISKL ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s -- - ... machine -mountpoint_8 0 880720 352416 40.0 5:411033.4 7:03 833.0 machine -mountpoint_9 FAILED --- machine -mountpoint_A FAILED --- machine -mountpoint_B 1 10 32 320.0 0:00 589.5 0:02 36.8 machine -mountpoint_C FAILED --- machine -mountpoint_D 1 679990 78272 11.5 1:061188.4 1:30 865.6 ... (brought to you by Amanda version 2.4.2p2) ---cut-off--- On the client-machine in /tmp/amanda I found no information leading to errors at sendsize.20020126004755.debug, nor at runtar.20020126???.debug but at amandad.20020126004937.debug: ---cut-on--- ... got packet: Amanda 2.4 REQ HANDLE 000-90B80708 SEQ 1012002305 ... sending ack: Amanda 2.4 ACK HANDLE 000-90B80708 SEQ 1012002305 ... amandad: got packet: Amanda 2.4 REQ HANDLE 000-78BA0708 SEQ 1012002304 amandad: received other packet, NAKing it addr: peer 192.168.1.15 dup 192.168.1.15, port: peer 854 dup 855 sending nack: Amanda 2.4 NAK HANDLE 000-78BA0708 SEQ 1012002304 ERROR amandad busy ... ... ... ---cut-off--- Why is amandad still busy? Or why ist the backup-server requesting a former sequenz? adTHANXvance Sascha
Re: Who uses Amanda?
John, thank you so much for processing the archives to extract the domain portions of folk's email addresses. I was hoping someone had the ability to do that, and appreciate the time and energy it took for you to do it. Thanks, again. -Kevin Zembower >>> [EMAIL PROTECTED] 01/25/02 07:07PM >>> >(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. ... Good answer :-). >Anyone have any guesses how many institutions and individuals are using >amanda? There are currently 1187 addresses on the amanda-users mailing list, and 561 on amanda-hackers. Running that all through uniq (and some other Perl magic), I came up with 1257 "domains" represented. Some caveats about that number. Not all sites running Amanda subscribe to the mailing lists, but not everybody subscribed to the list runs Amanda. Also, some of the addresses are clearly internal mailing lists, so the number of people who actually get the E-mail is certainly higher. Looking through the list, paying particular attention to .com's (since they are so much more important the rest of us peons :-), I see several names I recognize right away: 3com.com (3com) adp.com (ADP) attbi.com (AT&T) bbn.com (BBN) boeing.com(Boeing) corning.com (Corning) cypress.com (Cypress) daimlerchrysler.com (Chysler) dell.com (Dell) fedex.com (Federal Express) ge.com(General Electric) goodyear.com (GoodYear) harris.com(Harris) honeywell.com (Honeywell) hp.com(Hewlitt-Packard) ibm.com (IBM) informix.com (Informix) kodak.com (Kodak) mot.com (Motorola) nokia.com (Nokia) nsc.com (National Semiconductor) oracle.com(Oracle) philips.com (Phillips) redhat.com(Red Hat) ricoh.com (Ricoh) siemens.com (Siemans) sun.com (Sun) trw.com (TRW) valinux.com (VA Linux) xerox.com (Xerox) I'm sure there's a few billion dollars of worth floating around there, and I only looked at the U.S. .com entries. There are almost 500 international entries and another hundred .edu's (and if you don't think universities are in it for the money ... :-). >-Kevin Zembower John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Monthly-type backup
On 27 Jan 2002 at 12:37pm, Dan Smith wrote > Is there any way to do this elegantly? I was thinking I could also > setup 3 jobs and spread the disks out among them, but that would be > more work, and I'd like for amanda to organize the dumps so they best > fill out each tape. > dumpcycle=0, runtapes=3, chg-manual Then you get to play at being a tape robot! ;) -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: re-doing bad tapes
Ben Elliston wrote: > > On Friday, amdump wrote my dump images to a tape successfully, but a > subsequent amverify showed that the tape is defective. Now, my dumps have > been removed from the holding disk, but I don't think the images on the > tape are usable. > > Is there any way to re-do this backup onto a good tape or am I hosed? Do an amadmin amrmtape Mark the take so you don't re-use it. amlabel a new tape with the same Rerun the backup and you should get about the same levels as before. -- [EMAIL PROTECTED] - HMC UNIX Systems Manager My opinions are my own and probably don't represent anything anyway.
file too large, linux 2.4.2
Hi there list. I'm running amanda 2.4.2 on a RH7.1 box with 2.4.2-2. I recently read in the archives about making chunksizes 1Gb instead of 2 due to amanda tacking on stuff at the end, making those file too large. Too late, I'm afraid, as I'm trying to restore a set of files from a level0 of a sun box. I was able to get the archive off of the tape medium, and it's size is:2468577280. restore on the linux box in question fails with "File too large" (version 0.4b21). I was also able to get the file to the sun box it was backed up from, but ufsrestore complains that "Volume is not in dump format", which I assume is because it is a file made by dump, not ufsdump. So the question is: WHAT CAN I DO? Is there any way to get this directory extricated from this honking big 2 gig file? Second, I _thought_ that the 2.4 kernel was supposed to do away with the 2gb file size limitations. Am I misinformed? -- Matthew Boeckman(816) 777-2160 Manager - Systems Integration Saepio Technologies == == Public Notice as Required by Law: Any Use of This Product, in Any Manner Whatsoever, Will Increase the Amount of Disorder in the Universe. Although No Liability Is Implied Herein, the Consumer Is Warned That This Process Will Ultimately Lead to the Heat Death of the Universe.
Re: amrecover cannot find index for host
Kirk Strauser ([EMAIL PROTECTED]): > At 2002-01-25T19:16:22Z, "Jeremy Wadsack" <[EMAIL PROTECTED]> writes: >> We switch from tape- to hdd-based backup because hard drives are MUCH >> cheaper than tapes these days. I need to figure out the 2GB limit on the >> drives before I completely restart the system (separate problem.) > Ehhh? The good tape drives aren't particularly cheap, but the media is > practically throw-away money, especially for any shop larger than one > person. I mean, DDS-3 tapes are about $15, which translates to $0.625 per > GB. Add the fact that tapes are, by their nature, hot-swappable, and I > think you'd be hard-pressed to find a less-expensive and more-featureful HD > setup. Well, $0.625 / GB does not include the cost of the hardware. With a ten tape changer you *might* get 180GB of storage on "24GB" DDS-3 tapes. A 160GB hard drive is about $260. That's much lower cost than 10 DDS-2 tapes and the changer hardware. And it's significantly faster. It became impossible for us to backup 120GB of data in the off-hours of the server and the gtar process was using far too much resources on live web servers. -- Jeremy Wadsack Wadsack-Allen Digital Group
Re: nervous over amrecover
Albert Hopkins wrote: > > Testing a new amanda server install (RH Linux 7.2, amanda 2.4.2p2, DLT > 7000). I first ran a successful backup and then attempted to recover one > (small) file. The first time I extract I get > > amrecover: Can't read file header > extract_list - child returned non-zero status: 1 > > But then, without exiting I "add" and "extract" the file again. This > time it works. But this would have been pretty scary if this were a > real-life situation. I'm just wondering why this happened and what > could be done to prevent it. I had not ejected or rewound the tape > between backup and recovery. Could this have been the issue? You need to rewind the tape before doing the recover. -- [EMAIL PROTECTED] - HMC UNIX Systems Manager My opinions are my own and probably don't represent anything anyway.
Re: amandad
Davidson, Brian wrote: > > I have Amanda 2.4.3b1 installed on a intel computer running BSDI 4.1. I > have amanda in the /etc/services file and in the /etc/inetd.conf file. > amandad does not start when the computer is rebooted but no error messages > are reported in /var/log/messages. I can run amandad from the command line > as user "amanda" and it will creat the /tmp/amanda/debug.amanda directory > and debug file before timing out. That sounds like the appropriate behavior. Amandad doesn't run except when a connection comes in on port 10080. Running it by hand should do what you describe. -- [EMAIL PROTECTED] - HMC UNIX Systems Manager My opinions are my own and probably don't represent anything anyway.
Re: file too large, linux 2.4.2
On Mon, 28 Jan 2002 at 10:38am, Matthew Boeckman wrote > I'm running amanda 2.4.2 on a RH7.1 box with 2.4.2-2. I recently read > in the archives about making chunksizes 1Gb instead of 2 due to amanda > tacking on stuff at the end, making those file too large. Too late, I'm > afraid, as I'm trying to restore a set of files from a level0 of a sun > box. I was able to get the archive off of the tape medium, and it's size > is:2468577280. restore on the linux box in question fails with "File too > large" (version 0.4b21). I was also able to get the file to the sun box Newer versions of dump/restore should be compiled with large file support -- try upgrading. You could also compile the newest version yourself, adding in the appropriate flags. The kernel/glibc combo on RH7.1 *can* handle large files -- you just have to make sure that the app can. > it was backed up from, but ufsrestore complains that "Volume is not in > dump format", which I assume is because it is a file made by dump, not > ufsdump. No. For every filesystem, AMANDA runs the appropriate backup utility. Also, dumps are FS specific -- ext2 dump won't be able dump a UFS filesystem, and vice verse for ufsdump. Whether or not a particular "restore" can read another dump's format is often hit-and-miss. So, if that file was indeed from a Sun filesytem using "dump", it should indeed be a ufsdump archive. Perhaps it's getting truncated somewhere? > So the question is: WHAT CAN I DO? Is there any way to get this > directory extricated from this honking big 2 gig file? Second, I Your options are: 1) Use amrecover from the Sun box. This will automatically do everything in pipes and avoid the whole large file issue. 2) Upgrade dump/restore to a version compiled with large file support. 3) Pull the file off the tape with amrestore or dd, and pipe the output straight to restore, again avoiding a 2GB disk file. > _thought_ that the 2.4 kernel was supposed to do away with the 2gb file > size limitations. Am I misinformed? > As mentioned above, it's a kernel/glibc thing, *and* the app must be configured appropriately. IIRC, for whatever reason, the dump/restore shipped with RH7.1 wasn't. -- Joshua Baker-LePain Department of Biomedical Engineering Duke University
Re: Input/Output error
>Along those lines, here's what I see in dmesg after it finally fails: > >ide-tape: ht0: DSC timeout >ide-tape: ht0: position_tape failed in discard_pipeline() >ide-tape: ht0: DSC timeout >hdc: DMA disabled >hdc: ATAPI reset complete >ide-tape: ht0: I/O error, pc = 10, key = 2, asc = 4 ascq =1 >ide-tape: Couldn't write a filemark Well, those timeouts and resets can't be a good thing. I don't have the manuals for your specific drive (in fact I couldn't find it on the Seagate web site -- what kind of drive is it and what kind of tapes are you using?), but assuming it is roughly like the ones I have, the "key = 2, asc = 4 ascq =1" translates to "drive not ready, calibration in progress". Also, this confirms that some type of hardware error is happening. >> What happens if you try to amlabel one of your already labelled tapes? > >I get: > >rewinding, reading label DailySet101, tape is active >rewinding >tape not labeled > >so I guess it refused to do it. Is this what is supposed to happen? Sorry. I should have told you to use "-f" on amlabel to coerce it into rewriting the label even though the tape looks active. Now that I think about it some more, though, this can lead you down a path of other problems and is probably not a good idea to pursue. >> What happens if you run amcheck with the -w option ... >After switching to the tape Amanda wants to see next, I get: > >Amanda Tape Server Host Check >- >Tape DailySet100 is writable >Tape DailySet100 label ok >Server check took 4.565 seconds > >(brought to you by Amanda 2.4.2p2) > >Seems okay, is this what you expect? Yes, that looks normal. >I did this, and it seemed to work fine. No I/O errors, just nice little >messages from dd confirming reads and writes of so many blocks. I had 30 or >so writes in the script, and they all seemed to complete okay. Sigh. Oh, well. It's usually too much to ask for that hardware will misbehave with easy tests :-). >... Switched to the other IDE controller, still get the same behavior. That leaves the cable, the drive itself, or tremendously bad luck picking several tapes that all fail (my money is on the drive, or maybe the cable). >Steve Stanners John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: nervous over amrecover
>> ... The first time I extract I get >> >> amrecover: Can't read file header >> extract_list - child returned non-zero status: 1 >> >> But then, without exiting I "add" and "extract" the file again. This >> time it works. ... > >You need to rewind the tape before doing the recover. Chris is right. Here's a little more of what's going on behind the scenes: * When amdump (taper) completes, it leaves the tape where it left off (at the end). In theory you could use this to tack on some non-Amanda dump information to the tape, although that can be tricky. * Amidxtaped (called by amrecover) does **not** do any tape positioning before trying to read the image. This is so you can do the motion yourself on devices that fsf much faster than read/scan. The combination of these first two items probably gave you your initial error. * Amidxtaped *does* rewind the tape when it completes (even if there was an error), the theory being that if you need to do another restore from the same tape it will be in a known position to start from. This is probably why it worked the second time for you. Personally, I run a wrapper shell script around amdump and one of the things it does is unload any tape amdump writes to. So whenever I'm doing a restore, I have to mount the requested tape and therefor it's at a known position. I also always position the tape myself to do a restore (rewind followed by an appropriate fsf) because it's much faster on the types of drives I have. >> --a > [EMAIL PROTECTED] John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: amrecover cannot find index for host
>... A 160GB hard drive is about $260. That's much lower cost than >10 DDS-2 tapes and the changer hardware. ... Agreed, but there is a fundamental difference here and that's the critical failure path. If one of those 10 tapes fails, you should still be able to recover the majority of your data. If that single disk fails, you're screwed. I'm not disagreeing with the idea of using disk for backup. Just saying it takes some thought to make it a safe setup (tape does too, of course). >Jeremy Wadsack John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Monthly-type backup
On 27 Jan 2002, Dan Smith wrote: - I want to (once a month) do a backup of all disks. My tapes are not - large enough to fit the whole thing, so I need to spread it across 3 - tapes or so. I was thinking of setting up a 1-day dumpcycle with 3 - runspercycle and just do all three dumps on a saturday. I don't think - this is the best way to do it, since anything that has changed between - dumps will be backed up incremental to the 2nd or third run. - - Is there any way to do this elegantly? I was thinking I could also - setup 3 jobs and spread the disks out among them, but that would be - more work, and I'd like for amanda to organize the dumps so they best - fill out each tape. - - Does anyone have any advice? In amanda.conf define: define dumptype always-full { global compress none priority high dumpcycle 0 } Then in disklist put something like: server /volume { always-full } -- -- Stephen Carville UNIX and Network Administrator Ace Flood USA 310-342-3602 [EMAIL PROTECTED]
Re: FAILURE AND STRANGE DUMP SUMMARY:
>... suddenly 3 of 14 mountpoints on the same machine failed >for reasons I don't get. >... >machine /some/mountpoint_1 lev 1 FAILED [could not connect to machine] Some things to check: * See if there are left over amandad (or dumper) processes laying around still running from a previous failure. * Verify that the amandad service in inetd.conf/xinetd is set to "wait" (the default is "nowait"). * Make sure you don't have multiple cron jobs running at the same time. >Sascha John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
invalid password on windows host
When I run amcheck on the config for one of my windows machines I get the following: ERROR: lotus: [PC SHARE //vegas/avr access error: host down or invalid password?] I checked amandad.debug and found: running /usr/bin/smbclient vegas\\avr -E -U backup -W VPSA -c quit ERROR [PC SHARE //vegas/avr access error: host down or invalid password?] This would indicate to me that it is not picking up the usrid and password information stored in /etc/amandapass. The amandapass file has the correct permissions. I think I have the format correct. It looks like this: //vegas/avr userid%password NT_Domain Does anyone have any ideas? Thanks
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. I have no idea if Ace USA is "noteworthy" be we use Amanda to backup Oracle servers, Web servers, Samba servers and several development machines. Total data written to tape in a cycle is about 200 Gig. I originally picked it because the difference between the price of Amanda (free) and BackupExec licenses ($$$) paid for the 15 tape changers with a barcode reader. Now that I've been using it for a few months, I've discovered it is faster and more effecient of tape usage than BackupExec. The windows admin is actually jealous :-) What firewall problems are you having? -- -- Stephen Carville UNIX and Network Administrator Ace Flood USA 310-342-3602 [EMAIL PROTECTED]
bootstrapping
I did a few test with amanda, I tried both gnu tar and dump, I have 6 servers to dump, 3 Solaris and 3 Linux RedHat, some 50Gb, now I found what I think is the right configuration, bypassed the firewall and starting to dump the servers. And now I found the first problem I cannot easily solve. I am using a DDS4 tape, here is the def produced by tapetype: define tapetype DDS4-DAT { comment "just produced by tapetype program" length 16382 mbytes filemark 10 kbytes speed 1884 kps } and I?d like to use just one tape for each run, I've setup a 11Gb partition as holding disk and making dumps with dump: define dumptype sno-dump { global program "DUMP" comment "standard webserver dumped with dump" compress client best comprate 40,0 index yes priority medium holdingdisk yes record yes } The problem is that amanda (amanda-2.4.3b2-20020126) simply ignore the comprate parameter, I did a test run and I know that gzip does really a good work with those partition containing lots of html. I know that after the first run amanda will use the last one to obtain the estimates what I need is to do the initial full backups but it insists to keep the full size of the partition as estimate, in the first run I could just dump 3 partition out of 22: DUMP SUMMARY: DUMPER STATS TAPER STATS HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s --- -- - --- somehost /somewhere/ 0 1507440 618080 41.0 194:16 53.0 5:171948.1 somehost /somewhere/ 0 2372200 940672 39.7 270:57 57.9 8:021950.7 somehost /somewhere/ 0 10377660 7475360 72.0 111:261118.0 63:481952.8 So it looks like I used just 9gb out of 16 and all the other partition got better compression rates in the test phase than the first 2 that succeeded. I have 2 questions how can I convince amanda that a certain partition will be well compressed by gzip? Second I need to do the initial full backups, I am using: dumpcycle 9 # the number of days in the normal dump cycle runspercycle -1 # the number of amdump runs in dumpcycle days # (4 weeks * 5 amdump runs per week -- just weekdays) tapecycle 10 tapes # the number of tapes in rotation strategy standard may I run amdump 2 or 3 times the same day, so that all partitions get a level 0 dump or this can cause troubles with the dump cycle? Thanks in advance for the answer.
Re: bootstrapping
>define dumptype sno-dump { >... >comprate 40,0 >... >} > >The problem is that amanda (amanda-2.4.3b2-20020126) simply ignore the >comprate parameter ... I'll have to look into the code to see if it's really ignoring your values, but what you entered is not doing what you think. I assume what you're trying to say is that a compressed full dump will be 40% of the size of the original. If you look at the amanda(8) man page, you'll see the values for comprate are floating point and the defaults are 0.50 and 0.50, which means 50%. So you probably want: comprate 0.40,0.40 Your value of 40 is telling Amanda the resulting compressed image will be 40 times larger than the original. No wonder it's ignoring you :-). Since this is only for getting started (as you noted, Amanda will use its own history after the initial run), you might just ignore setting comprate since 50% is not all the different than 40%. >may I run amdump 2 or 3 times the same day, so that all partitions get a >level 0 dump ... Yes. John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]