Error in compiling Amanda
Hi, I'm using the pre-build package of Amanda from Debian 3.1 until now. But I want to try to improve my DLT-Streamer by feeding him with a bigger block-size. For this, I tried to compile Amanda direct from source with "--with-maxtapeblocksize 64". I tried this with Amanda 2.5.1p2 and will only get this error after configure and make: creating sendbackup if gcc -DHAVE_CONFIG_H -I. -I. -I../config -I../common-src -I../common-src -I../amandad-src -Wall -W -Wparentheses -Wmissing-prototypes -Wstrict-prototypes -Wmissing-declarations -Wformat -Wsign-compare -Wfloat-equal -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -MT sendsize.o -MD -MP -MF ".deps/sendsize.Tpo" -c -o sendsize.o sendsize.c; \ then mv -f ".deps/sendsize.Tpo" ".deps/sendsize.Po"; else rm -f ".deps/sendsize.Tpo"; exit 1; fi sendsize.c: In Funktion »getsize_dump«: sendsize.c:1285: error: `is_rundump' undeclared (first use in this function) sendsize.c:1285: error: (Each undeclared identifier is reported only once sendsize.c:1285: error: for each function it appears in.) make[1]: *** [sendsize.o] Fehler 1 make[1]: Leaving directory `/root/src/amanda-2.5.1p2/client-src' make: *** [all-recursive] Fehler 1 What did I made wrong? Thanks, Thoddi
Re: How to restore correct ownerships?
Moin Jon, > The difference is that the backup programs are likely to > run unattended, or by an operator and the result is a collection > or archive of all the files, possibly on tape. To get root > privlege during backup, certain programs are "setuid'ed" to > root. Thus they can, and should be run by the backup user > despite needing root privlege for certain operations. Yes, I know that Amanda setuid'ed himself to run some tasks with higher privileges. But why he didn't setuid also while restore short before the chown? When restoring as root, I had to enter a " root"-line to the amandahosts-file. Until now there only must be a " backup"-line for normal backup-tasks, because he uses setuid. Could'nt this new root-entry be a security-hole now? On the other hand: I can just add this line manually before each restore. And if somebody cracks my server, then I have other problems than this :-) The root-entry didn't listen on external-interfaces, so what?! :-) > For a few months, until my game was discovered, my bill dropped > to about $300/month. Yes, I know such playing with quota. In the beginning of my CS-studies our quota-software could'nt count the quota when moving files to another harddisk. So another student on a different disk installs a directory for another one. And after that, you could get space only limited by the space of the other disk(s) :-) But: Paying for disk-space as you? How long is this away? :-)) -- Bis denne, Thoddi
Re: How to restore correct ownerships?
Moin Paul, >> I just needed to restore some deleted files via amrecover and it works >> well. But I noticed that the files will be restored with the ownership >> "backup:backup" instead of the original "thoddi:users". > You should restore as user "root", not as user "backup". Well, why this? Another one on this list told me that all Amanda-operations should NOT be done as root, but as the backup-user. After adding extra root-entries to the amndahosts, it now works, thanks. But I don't understand why I have to backup as "backup" and must restore as "root"... -- Bis denne, Thoddi
How to restore correct ownerships?
Hello, I just needed to restore some deleted files via amrecover and it works well. But I noticed that the files will be restored with the ownership "backup:backup" instead of the original "thoddi:users". OK, I can make a chown over these 10 files manually, but I have backup-points, where the files and directories have many different GIDs... So a larger correct recorery will be tricky or impossible. Can Amanda save and restore the ownerships or is that not possible? Thanks. -- Bis denne, <° >< Thoddi
Collecting dumps bevore tape-writing?
Hello, in the "normal" way, amanda collects the dumps on the holding disk and write them to tape, while amanda dumps another directory in the background and so on. But after installing a new streamer, my computer is too slow for this. I can only dump to the holding disk OR writing data to tape at one time, not both together (the streamer is too fast, the disk too slow). Yes, a hardware-update is on the way :-)) Is there a way to tell amanda, that he must not dump and tape-writing in parallel, but one step after the other? Thank's. -- Bis denne, <° >< Thoddi
How uses Amada a tapechanger?
Hello, until now I'm using a single DDS-3 DAT-streamer with Amanda and everything works well. But it is annoying for me, to change the tape every day. Sometime I'll forget this or I'm not at home. So it is possible that a lot of backups are collecting on the holding disk. I could buy a used "ADIC FastStor DLT 4000"-tapechanger (For 7 cartrigdes, each 20 GB native capacity). Is it possible to use this tapechanger together with Amanda? After reaging the docs I also didn't understand, HOW Amanda could use a tapechanger: Is it possible, that I could say, that he should use 2 tapes per tape-run, to save 40 GB on 2 tapes? Or is a tapechanger only 7 single tapes for Amanda? And the only difference is, that a machine automatically change the tapes every day (instead of me)? -- Bis denne, <° >< Thoddi
Re: Problems using amrecover
Moin Paul, > Your localhost is actually called "schatten", and I guess that a > dns lookup of "schatten" does not result in "localhost", nor in 127.0.0.1. Hmm... you're right :-) > Don't use "localhost" in disklist. It will bit you. > If you don't believe me, just just got bitten... OK, I begin to understand... Any chance to resolve this situation here? I tried to re-manage all here: My "domain" at home is called "magnus" (instead of something else, part of history, but this shouldn't affect the problem, because DNS works well with that :->), so nslookup will say: Name: schatten.magnus Address: 10.10.10.1 1.10.10.10.in-addr.arpa name = schatten.magnus. So, correct until here. /etc/amandahosts: schatten.magnus root schatten root disklist: schatten.magnus /home/thoddi comp-user-tar amcheck without any errors after renaming the "localhost"-dir in "curinfo" and "index" (because they are the same, just hostnames changed :->) But: schatten:~# amrecover Weekly -s schatten.magnus AMRECOVER Version 2.4.2p2. Contacting server on schatten.magnus ... 220 schatten AMANDA index server (2.4.2p2) ready. 200 Access OK Setting restore date to today (2004-05-27) 200 Working date set to 2004-05-27. 200 Working date set to 2004-05-27. amrecover: Unexpected server end of file <#ß=(%*'-GRMPF-§"§$")='_:>:-) Why he told me "200 Access OK" first and will fail an few lines after? Syslog only said: May 27 17:06:16 schatten amindexd[11901]: connect from 10.10.10.1 I first thougt about iptables blocking the connection, but the packetfilter is working correct. Is the backup gone away? No way to recover it? -- Bis denne, Thoddi
Problems using amrecover
Hello, I got some "funny" problems using amrecover: schatten:/etc# amrecover Weekly AMRECOVER Version 2.4.2p2. Contacting server on localhost ... 220 schatten AMANDA index server (2.4.2p2) ready. 200 Access OK Setting restore date to today (2004-05-27) 200 Working date set to 2004-05-27. 200 Working date set to 2004-05-27. amrecover: Unexpected server end of file Both server and client are on the same machine, a normal backup is going without any problem. As amrecover said, everything is "localhost" :-) I found this note in the FAQ: http://amanda.sourceforge.net/fom-serve/cache/173.html I tried all in there, I can ping to and from localhost :-) and nslookup tell me, that localhost is really 127.0.0.1 and vice versa :-) Because a normal backup is working, I think that access-rights and so on must be correct. So: Please HELP... The system I'm running amanda is a Debian 3.0 (Woody) using the amanda-packets shipped with this distribution. -- Bis denne, <° >< Thoddi
Connection to my floppy?
Hello, when using Amanda I have this problem: with EVERY amanda-command (i.e. amcheck or amdump), first I see a connection to my floppy disk. When no disk is inserted, my console is filled with error-messages ("end_request: I/O error, dev 02:00 (floppy), sector 0") and after some tries, the amanca-command started working. When using amdump, amanda do this before every new file on the holding disk, after that, he copies the data to the holding disk and started filling the tape as normal. I gave him a floppy-disk to eat (unformatted and formatted with FAT, Minix and so on), but I can see no change on the floppy-disk after that. After making a loud noise in my floppy, Amanda works perfectly :-) My question: What the hell will amanda do on my floppy? I didn't think that we really tries to store data on 1.44 MB :-)) How can I stop this behaviour. Thanks. -- Bis denne, <° >< Thoddi