amlabel
hi all, i 'm a Amanda Newbie using amanda-2.4.2, executing: su amanda amlabel -f amanda.conf DailySet1-000 getting the error message: amlabel: amlabel: cannot execute binary file I don't know what i did wrong... i had compiled all those nice stuff on my Solaris8i386 maschine. ich had create: /usr/local/etc/amanda/DailySet1 stored there: -rw-r--r-- 1 amanda backup 19892 Mar 15 09:52 amanda.conf drwxr-xr-x 2 amanda backup 512 Mar 13 16:38 config -rw-r--r-- 1 amanda backup 14498 Mar 14 11:10 config.site -rw-r--r-- 1 amanda backup 144 Mar 15 11:23 disklist -rw-r--r-- 1 amanda backup 161 Mar 15 09:54 tapelist is this right too? many thanx fore some good answeres! Bernd Zimmermann
Re: amlabel
You put the directory name as the config name: su amanda amlabel -f DailySet1 DailySet1-000 ^ On Thu, 15 Mar 2001, Bernd Zimmermann wrote: > hi all, > i 'm a Amanda Newbie using amanda-2.4.2, executing: > su amanda amlabel -f amanda.conf DailySet1-000 > getting the error message: > amlabel: amlabel: cannot execute binary file > > I don't know what i did wrong... > i had compiled all those nice stuff on my Solaris8i386 maschine. > ich had create: > /usr/local/etc/amanda/DailySet1 > stored there: > -rw-r--r-- 1 amanda backup 19892 Mar 15 09:52 amanda.conf > drwxr-xr-x 2 amanda backup 512 Mar 13 16:38 config > -rw-r--r-- 1 amanda backup 14498 Mar 14 11:10 config.site > -rw-r--r-- 1 amanda backup 144 Mar 15 11:23 disklist > -rw-r--r-- 1 amanda backup 161 Mar 15 09:54 tapelist > is this right too? > many thanx fore some good answeres! > Bernd Zimmermann > > -- Jason Hollinden SMG Systems Admin
AW: amlabel
mmh, ich tryed this too. but there ist is still the same message. >You put the directory name as the config name: > >su amanda amlabel -f DailySet1 DailySet1-000 > ^ > >On Thu, 15 Mar 2001, Bernd Zimmermann wrote: > >> hi all, >> i 'm a Amanda Newbie using amanda-2.4.2, executing: >> su amanda amlabel -f amanda.conf DailySet1-000 >> getting the error message: >> amlabel: amlabel: cannot execute binary file >> >> I don't know what i did wrong... >> i had compiled all those nice stuff on my Solaris8i386 maschine. >> ich had create: >> /usr/local/etc/amanda/DailySet1 >> stored there: >> -rw-r--r-- 1 amanda backup 19892 Mar 15 09:52 amanda.conf >> drwxr-xr-x 2 amanda backup 512 Mar 13 16:38 config >> -rw-r--r-- 1 amanda backup 14498 Mar 14 11:10 config.site >> -rw-r--r-- 1 amanda backup 144 Mar 15 11:23 disklist >> -rw-r--r-- 1 amanda backup 161 Mar 15 09:54 tapelist >> is this right too? >> many thanx fore some good answeres! >> Bernd Zimmermann
Re: amlabel
H > i 'm a Amanda Newbie using amanda-2.4.2, executing: > su amanda amlabel -f amanda.conf DailySet1-000 > getting the error message: > amlabel: amlabel: cannot execute binary file > > I don't know what i did wrong... It might be worth BECOMING the amanda user and doing it. I always find "su user (some command)" tends to get things confused... DSL -- The Question Is: "Can AUUG augment LinuxSA?" (maybe it would be easier to ask Cornelia "who is the weakest link?")
Re: Re: Best config for a single large root partition
Hi Jens, >Yura Pismerov <[EMAIL PROTECTED]> writes: > >> You can't setup a separate exclude list for each entry in disklist file. >> In this particular case you just need to put ./home and ./var into >> exclude file and make 3 separate entries for /, /home and /var in the >> disklist. I assume you don't have ./home and ./var under /home and /var >> (i.e /home/var or /var/home directories). > >2.4.2 makes it possible to define dump types directly in the >disklist. And since dump types can define the filename for an exclude >file you can have seperate exclude files for each disk entry (of >course you could also define the dump types in amanda.conf and use >different dump types for each disk entry). > This is an interesting new feature for me. >This is very useful if you change the disklist automatically (I use it >for splitting up a ~300 GB filesystem so every disk entry will fit on >a 35 GB DLT tape). > >In the case of splitting / into /, /home and /var I would define a >dumptype directly in the disklist for / with an exclude file >containing ./home and ./var as suggested above and use a "normal" dump >type for /home and /var. > Some things are unclear to me. 01 Because of dump don't support exclude file lists/patterns, I have to use gtar, if I want to benefit from this feature? 02 If I create an exclude file for the '/home' disklist entry, what about the files in '/' ? If I include an entry like '/*' in those exclude list, '/home' won't be tar'ed too? 03 In the amanda man page I read about the exclude amanda.conf parameter: exclude [list] "string" Default: none. The string is passed as a file name value to the GNU tar --exclude or --exclude-from option. In the case of exclude list, the file must be available on the backup client host. and from tar --help I read about the --exclude parameter: --exclude=PATTERNexclude files, given as a globbing PATTERN -X, --exclude-from=FILE exclude globbing patterns listed in FILE 01 My assumption is, that the amanda exclude "string" (not using the list option) is passed to the tar parameter that expects the globbing file PATTERN? 02 If yes, can I define in the local dumptype an exclude PATTERN for the appropriate disk entry, so I don't have to create and maintain a corresponding exclude list file at the client side? Which will mean, can I maintain all exclude PATTERNS in the disklist file? >/Jens >-- >Jens Bech Madsen >The Stibo Group, Denmark mike
Re: Need help understaning dump output
On Wed, 14 Mar 2001, John R. Jackson wrote: > >... I am encrypting the backups with gnupg and using tar. ... > > Huh? How are you doing the encryption? Amanda doesn't support that. > Check out this url: http://security.uchicago.edu/tools/gpg-amanda Pretty easy to get going. Just uses a wraper on the client. > >When I run amdump I get > >the error "dump larger then tape, but cannot incremental dump new disk" > > This says the estimated dump size Amanda got is larger than 3274 mbytes > (the "length" value in your tapetype). > So what I am hearing is that amanda can backup no partition larger then the size of the tape drive + whatever compression is in use. > >Can amanda span a dump across multiple tapes? ... > > Not yet. I madly trying to work on this in the five spare minutes I get > a day as I now have file systems that won't fit, even with compression. > OK. One last question how do you specify to use hardware or sofware compression? Thankyou for your time. Andrew Hall
Re: SGIs not backing up
On Wed, 14 Mar 2001 17:54:50 -0500 "John R. Jackson" <[EMAIL PROTECTED]> wrote: > >I've taken over the amanda backups of a number of machines. They seem > >to be working now except that 3 machines fail with this error message > >like this: > > FAIL planner glenord /dev/dsk/dks1d5s7 0 [Request to glenord failed: > > timeout waiting for ACK] > > Based on that message, your server appears to be running the main CVS > code, otherwise known as "2.5" or the development (i.e. "expermental") > branch. Is that correct and did you realize that? It is correct. I was unaware of this. You see I am trying to fix this for a network that I am not familiar with. And the person who set it up has left. The server is running "Amanda-2.5.0-19990816" and the SGI clients "Amanda-2.4.2-19990203". > If you don't want > to be on the bleeding edge you might consider going back to 2.4.2p1, > the latest stable release. I certainly don't want to be on the bleeding edge, but I'm not sure how much effort will be required to reconfigure the server and all the Sun clients to the earlier version. I mean I have the configuration file amanda.conf, so will there be much effort involved in getting 2.4.2p1, installing the client and server side, and then could I just use the old amanda.conf file with the tapelist and disklist files and it should continue from where it left off? ie I just replace the the 2.5 binaries with the older 2.4.2 ones? > > I've seen a timeout problem where clients were not at 2.5.0 and the > server was. It has to do with the new protocol engine in the security > code in 2.5, but I didn't bother working on it at the time (although I > seem to recall posting to amanda-hackers, with no reply). > > >... The fix is probably fairly obvious ... > > Actually, I suspect the fix is going to require some reasonably > good coding, C level and protocol debugging experience on your part. I'm afraid I can't spare the time to do this. > You might search the amanda-hackers archive for my post and see if that > gets you started. I'll have alook. Thanks for your time, Ross > > >Is it got something to do with the fact that it has xfs type > >filesystems and an efs filesystem? Do I have to tell amanada (somehow) > >about this? > > No, this isn't the problem and no, you don't have tell Amanda about this. > It figures it out for itself. > > >Ross > > John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] -- Ross Macintyre Heriot-Watt University [EMAIL PROTECTED]
Re: FATAL: data write: File too large
On 15 Mar 2001, Alexandre Oliva wrote: > > and is: > > chunksize 2gb > > the right way to limit it? > > Nope. Use 2000mb. That's a little bit less than 2Gb, so it won't > bump into the limit. Confusing, but understood. I will try that, thanks. Mark -- http://www.mchang.org/ http://decss.zoy.org/
amanda mail report size
Ever since I rebuilt Amanda and moved it over to another box, I have noticed that the email report size wise is huge (9 MB). Didn't recall the report's to be that big on the old amanda instance I had running. Had an amanda alias pointing to my corp email address, but now that the message size is so big, it's getting rejected at the mail server (not under my control) I've since pointed the amanda alias to my local account(plenty of control ;)) and get the reports fine... just takes forever to load the report up (mutt) ;) Is there any options on amreport, like setting a verbose level? I'm really only interested in dump summary. Thanks, -Ben
AW: amlabel
High: The error message indicates that the file amlabel is not a true executable program file even though it's file attributes contain 'x' for executable. Perhaps there is an additional file amlabel in the PATH which is not the program amlabel. Regards K.H. Timmesfeld -- == Dr. K.H. Timmesfeld Tel.: 06431 -4040 (Sekretariat) IDAS GmbH -40439 (Durchwahl) Holzheimer Strasse 96 Fax:-40410 D-65549 Limburg email: [EMAIL PROTECTED] privat: Im Feldchen 2 Tel. und Fax: 06431-41271 D-65549 Limburg
Re: amanda mail report size
Ben Hyatt wrote: > > Ever since I rebuilt Amanda and moved it over to another box, I have noticed > that > the email report size wise is huge (9 MB). Doesn't this mean that there is something wrong with backup ? > > Didn't recall the report's to be that big on the old amanda instance I had > running. > Had an amanda alias pointing to my corp email address, but now that the > message size is so big, > it's getting rejected at the mail server (not under my control) > > I've since pointed the amanda alias to my local account(plenty of control > ;)) and get the reports fine... just takes forever to load the report up > (mutt) ;) > > Is there any options on amreport, like setting a verbose level? I'm really > only interested > in dump summary. > > Thanks, > > -Ben -- Yuri Pismerov, TUCOWS.COM INC. (416) 535-0123 ext. 1352 S/MIME Cryptographic Signature
dump returned 3
Hi there, I'm getting some strange reports in my dump summary: pluto /dev/ida/c0d0p8 lev 0 FAILED [/sbin/dump returned 3] pluto /dev/ida/c0d0p8 lev 0 FAILED [dump to tape failed] And here is a snippet of dump's output for this device... | DUMP: 71.14% done, finished in 2:50 | DUMP: 72.60% done, finished in 2:40 | DUMP: 72.94% done, finished in 2:39 ? DUMP: bread: lseek fails ? DUMP: bread: lseek fails ? DUMP: bread: lseek fails ? DUMP: bread: lseek fails ? DUMP: bread: lseek fails ? DUMP: bread: lseek fails ? DUMP: bread: lseek fails ? DUMP: bread: lseek fails ? DUMP: bread: lseek fails ? DUMP: short read error from /dev/ida/c0d0p8: [block -200696]: count=4096, got=0 ? DUMP: DUMP: bread: lseek fails ? DUMP: short read error from /dev/ida/c0d0p8: [block -1451271504]: count=4096, got=0 ? DUMP: DUMP: bread: lseek fails ? bread: lseek2 fails! ? DUMP: short read error from /dev/ida/c0d0p8: [sector -200696]: count=512, got=0 Does anyone know what the dump returned 3 error means and what can be done about it? Finally I'm getting the error theakston2 /dev/rd/c0d1p1 lev 0 FAILED [Request to theakston2 timed out.] Which seems pretty strange seeing as though amcheck was fine about it! Any ideas? -- Moneyextra Limited is an independent intermediary regulated by the Personal Investment Authority. The PIA does not regulate mortgages, current or deposit accounts, credit cards, loans or general and medical insurance. Moneyextra Limited, One Temple Back East, Temple Quay, Bristol BS1 6DX
RE: amanda mail report size
> Doesn't this mean that there is something wrong with backup ? I have 8 samba shares which get backed up - due to open files I get a lot of errors from samba. I added 2 samba shares to the disklist (previous instance had 6 samba shares defined) my previous reports were never this big. -Ben > Yuri Pismerov, TUCOWS.COM INC.(416) 535-0123 ext. 1352
Re: dump returned 3
On Thu, Mar 15, 2001 at 05:57:09PM +, Raymond Bramwell wrote: It is possible you're trying to dump a corrupted filesystem. If the filesystem checks out OK, could it be you're trying to dump an active filesystem? Dump gets down and dirty with the details of the filesystem, and if the filesystem has changed between the time dump reads a reference and the time it picks up the referenced data, it may wind up interpreting what's now, for example, just plain data, as further filesystem references (block numbers, offsets, and so on). If that only happens once in a while, perhaps it is tolerable. Alternately, you might consider trying GNU tar. This is a FAQ on this list. The "timed out" is disquieting. I've seen such errors mostly when there is failing hardware. Failing hardware, of course, is a quick recipe for filesystem corruption. You might want to take a look at your logs, and see if you're encountering "timed out" at other times. I use a scanner that looks for this message on all machines several times a day, and emails me the result, if it finds any new ones. Most cost-effective way I've found yet to catch failing hard drives before they do much damage. If this was one of my machines I'd try to find a time I could bring it down very soon, check the filesystems, and do a complete read-only test of the disk surface. > Hi there, I'm getting some strange reports in my dump summary: > > pluto /dev/ida/c0d0p8 lev 0 FAILED [/sbin/dump returned 3] > pluto /dev/ida/c0d0p8 lev 0 FAILED [dump to tape failed] > > And here is a snippet of dump's output for this device... > > | DUMP: 71.14% done, finished in 2:50 > | DUMP: 72.60% done, finished in 2:40 > | DUMP: 72.94% done, finished in 2:39 > ? DUMP: bread: lseek fails > ? DUMP: bread: lseek fails > ? DUMP: bread: lseek fails > ? DUMP: bread: lseek fails > ? DUMP: bread: lseek fails > ? DUMP: bread: lseek fails > ? DUMP: bread: lseek fails > ? DUMP: bread: lseek fails > ? DUMP: bread: lseek fails > ? DUMP: short read error from /dev/ida/c0d0p8: [block -200696]: > count=4096, got=0 > ? DUMP: DUMP: bread: lseek fails > ? DUMP: short read error from /dev/ida/c0d0p8: [block -1451271504]: > count=4096, got=0 > ? DUMP: DUMP: bread: lseek fails > ? bread: lseek2 fails! > ? DUMP: short read error from /dev/ida/c0d0p8: [sector -200696]: > count=512, got=0 > > Does anyone know what the dump returned 3 error means and what can > be done about it? > > Finally I'm getting the error > > theakston2 /dev/rd/c0d1p1 lev 0 FAILED [Request to theakston2 timed > out.] > > Which seems pretty strange seeing as though amcheck was fine about it! > Any ideas? > > -- > Moneyextra Limited is an independent intermediary regulated by the > Personal > Investment Authority. The PIA does not regulate mortgages, current or > deposit accounts, credit cards, loans or general and medical insurance. > Moneyextra Limited, One Temple Back East, Temple Quay, Bristol BS1 6DX > -- - Dan Wilder <[EMAIL PROTECTED]> Technical Manager & Correspondent SSC, Inc. P.O. Box 55549 Phone: 206-782-7733 x123 Seattle, WA 98155-0549 URLhttp://www.linuxjournal.com/ -
Re: amanda mail report size
Ben Hyatt wrote: > > > Doesn't this mean that there is something wrong with backup ? > > I have 8 samba shares which get backed up - due to open files I get a lot > of errors from samba. > > I added 2 samba shares to the disklist (previous instance had 6 samba shares > defined) my previous reports were never this big. I don't think you can control it on amanda level since the messages are generated by smbclient. Patch the smbclient ? > > -Ben > > > Yuri Pismerov, TUCOWS.COM INC.(416) 535-0123 ext. 1352 -- Yuri Pismerov, TUCOWS.COM INC. (416) 535-0123 ext. 1352 S/MIME Cryptographic Signature
RE: FATAL: data write: File too large
> > > > Nope. Use 2000mb. That's a little bit less than 2Gb, so it won't > > bump into the limit. > > Confusing, but understood. I will try that, thanks. I think the amanda man page discusses this a bit, or else it's in the default amanda.conf. To allow for header information on the dump file, etc., you can't set chunksize to exactly 2GB (= 2048MB); it has to be a bit less. 2000MB is a nice, round, easy-to-remember number. Carey
RE: amanda mail report size
On Thu, 15 Mar 2001, Ben Hyatt wrote:
> I have 8 samba shares which get backed up - due to open files I get a lot
> of errors from samba.
>
> I added 2 samba shares to the disklist (previous instance had 6 samba shares
> defined) my previous reports were never this big.
you can patch amanda to be less verbose when samba (or other) errors are
encountered. you only need to do this on the samba client that is the
host for the smbclient backups. that is, if your disklist looks like:
host1 //pchost/driveC comp-user-tar
host1 //pchost/driveD comp-user-tar
host2 / comp-user
you only need to patch amanda for host1.
there are two patches you can apply that affect two different aspects of
the smbclient backup.
1) client-src/sendbackup-gnutar.c
look for this code segment:
#ifdef IGNORE_SMBCLIENT_ERRORS
/* This will cause amanda to ignore real errors, but that may be
* unavoidable when you're backing up system disks. It seems to be
* a safe thing to do if you know what you're doing. */
{ DMP_NORMAL, "^ERRDOS - ERRbadshare opening remote file", 1},
{ DMP_NORMAL, "^ERRDOS - ERRbadfile opening remote file", 1},
{ DMP_NORMAL, "^ERRDOS - ERRnoaccess opening remote file", 1},
{ DMP_NORMAL, "^ERRSRV - ERRaccess setting attributes on file", 1},
{ DMP_NORMAL, "^ERRDOS - ERRnoaccess setting attributes on file", 1},
#endif
if you define IGNORE_SMBCLIENT_ERRORS then these smbclient error messages
will be treated as DMP_NORMAL and they will not cause a full error report
to be generated (ie. amanda will output all messages from the dump if any
non-DMP_NORMAL lines are detected. this is the problem you are seeing.).
you can define IGNORE_SMBCLIENT_ERRORS in your configure stage by, for
example, putting the following in config.site:
CFLAGS=-DIGNORE_SMBCLIENT_ERRORS
2) if you don't care to see any DMP_NORMAL lines in the report, only
DMP_STRANGE lines, look at client-src/sendbackup.c for this code:
typ = parse_dumpline(str);
switch(typ) {
case DMP_NORMAL:
case DMP_SIZE:
startchr = '|';
break;
default:
case DMP_STRANGE:
startchr = '?';
break;
}
fprintf(stderr, "%c %s\n", startchr, str);
and add a "return" statement after the "case DMP_NORMAL" line:
case DMP_NORMAL:
return;
so that the fprintf is not called for DMP_NORMAL lines. the downside to
this is that you will not see any of the possibly useful DMP_NORMAL
information when there is a DMP_STRANGE error. this could make debugging
the problem more difficult. it would be useful if this was a run-time
configurable feature.
--
Todd Pfaff \ Email: [EMAIL PROTECTED]
Computing and Information Services \ Voice: (905) 525-9140 x22920
ABB 132 \ FAX: (905) 528-3773
McMaster University \
Hamilton, Ontario, Canada L8S 4M1 \
amrecover vs. amrestore
Can someone give me an idea of amrecover vs. amrestore vs. just getting the tars off the tape manually? I just ran my first "test" backup and didn't turn on indexing, so I can't use amrecover (actually, can't test it). I hear amrecover is pretty neat with the file-browsing... Mark -- http://www.mchang.org/ http://decss.zoy.org/
amandad stand alone????
Can anyone tell me if it is possible to run amanda in a stand alone config (i.e. not via inetd or brethren)? I am trying to set up amanda on a group of external servers. Because of the limited number of services offered we don't/won't want to run inetd. What we want to do is be able to start and stop amandad only during the backup window. This way we can wrap the amandad startup in a script that only enables the related ipchains rules during the window. If amandad doesn't have a daemon option can anyone provide ideas on what might be options other than inetd/xinted? Thanx (hopefully my objectives make sense) -Dustin
Re: amandad stand alone????
On Thu, 15 Mar 2001, Duncan, Dustin wrote: > Can anyone tell me if it is possible to run amanda in a stand alone config > (i.e. not via inetd or brethren)? i'm not sure if this is possible. if not... > I am trying to set up amanda on a group of external servers. Because > of the limited number of services offered we don't/won't want to run > inetd. What we want to do is be able to start and stop amandad only > during the backup window. This way we can wrap the amandad startup > in a script that only enables the related ipchains rules during the > window. > > If amandad doesn't have a daemon option can anyone provide ideas on > what might be options other than inetd/xinted? tcpserver might be a viable alternative. http://cr.yp.to/ucspi-tcp/tcpserver.html http://cr.yp.to/ucspi-tcp.html -tcl.
exclude list format
Hello, I have a 25 gig partition that I need to backup. My tape drive is only a 4/8 gig drive. My thought was to create an exclude list and run the backup in two steps on two tapes. I am only using 12 of the 25 gigs. My dumptype has the line: otions exclude-list "/etc/amanda/Backup-Test/exclude.sda5" that file exists on the client and is owned by user and group backup. The contents are: /var/FTP /var/HOMES The problem is that even though I am telling it to exclude these files AND use tar it still creates the tarball with those directories which causes the dump to fail. What is the syntax of the exclude file. Does anyone know of a GOOD place for amanda docs. All I can find says use fearture A but gives no way of how to use feature A. For example the man page talks about the exclude file but does not cover the format of the exclude file. Why does amcheck not find these errors? Thanks again for your time. Andrew Hall
exclude list format GOT IT
ahhh it works on files not directories. Sorry to send an unnecessary mail. Andrew -- Forwarded message -- Date: Thu, 15 Mar 2001 16:18:30 + (GMT) From: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: exclude list format Hello, I have a 25 gig partition that I need to backup. My tape drive is only a 4/8 gig drive. My thought was to create an exclude list and run the backup in two steps on two tapes. I am only using 12 of the 25 gigs. My dumptype has the line: otions exclude-list "/etc/amanda/Backup-Test/exclude.sda5" that file exists on the client and is owned by user and group backup. The contents are: /var/FTP /var/HOMES The problem is that even though I am telling it to exclude these files AND use tar it still creates the tarball with those directories which causes the dump to fail. What is the syntax of the exclude file. Does anyone know of a GOOD place for amanda docs. All I can find says use fearture A but gives no way of how to use feature A. For example the man page talks about the exclude file but does not cover the format of the exclude file. Why does amcheck not find these errors? Thanks again for your time. Andrew Hall
DELL 120T changer on RH 6.2
Hello All, I use Amanda 2.4.2p1 on RH Linux 6.2. we have Dell PowerVault 120T, and the documentation says that it has two SCSI devices - one for tape device and second for tape changer. Taper is running as /dev/nst0, seems to be OK - at least it make some noise . But I couldn't find any way to get the changer working. I used /dev/sch0, /dev/sg*(means sg0-sg3) devices, with no respond. Last time I set it to /dev/sga, and it doesn't give error messages, but I can't force it to do any operation with tapes. Does anybody have PV120T changer working under RH linux ? Any help and suggestions are welcome. Murat Dymov _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com
Re: Need help understaning dump output
>So what I am hearing is that amanda can backup no partition larger then >the size of the tape drive + whatever compression is in use. Correct, for the moment. I'm trying to fix that (literally) as I'm writing this. >OK. One last question how do you specify to use hardware or sofware >compression? If you use software compression, then you either pick a device name or use whatever method your OS provides to disable hardware compression, then use the "compress" dumptype options in amanda.conf. If you use hardware compression, you use "compress none" in amanda.conf and turn on hardware compression with the device name or whatever. When you use hardware compression, you need to lie to Amanda about the tape size. I typically crank it up about 20-30% (don't double it like the vendors say you can -- they live in a dream world :-). >Andrew Hall John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Failed queue message - Amanda 2.4.2p1 - Investigation results
Further investigation results Hi John, FYI and also for others. In Solaris 8 the raw devices are set to crw--- root sys . ufsdump does a check on the real user id and will not let you at the raw device if you are not really root even though the link from the /dev directory is setuid. Chmod to allow crw-r--r-- allowed this to work. I will also add amanda to the sys group and re-chmod to crw-r-. This should provide a little more security (as these machines are on the net). Many thanks for your assistance. Regards "John R. Jackson" wrote: > > >... Don't you ever sleep? > > I'm beginning to think it's a waste of time :-). > > >Sorry for appearing dumb but sendsize*debug? What it is? > > It's a file in /tmp/amanda on the client. Lots of good stuff going on > in there. > > >ERROR: lagoon.ind.tansu.com.au: [can not read/write /etc/dumpdates: > >Permission denied] > >... > >[amanda@lagoon]% ls -al /etc/dumpdates > >-rw-rw-r-- 1 root sys0 Feb 9 14:37 /etc/dumpdates > > Inetd fires up amandad as some particular user on this client. > What groups is that user a member of? In particular, is it a member of > group "sys"? If not, you either need to add it to group "sys" or change > the group on /etc/dumpdates to a group it is a member of (the latter is > the usual course of action). > > >David Logan > > John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] -- ___ David Logan 102 Cameron Street Broadband e-Lab Launceston Tas 7250 Australia Ph : (03) 6323 2667 Fax: (03) 6323 2666 Mob: (0419) 890 763
Re: amlabel
> su amanda amlabel -f amanda.conf DailySet1-000 >getting the error message: > amlabel: amlabel: cannot execute binary file > >I don't know what i did wrong... >i had compiled all those nice stuff on my Solaris8i386 maschine. I'm not sure if it was just a typo when you wrote your letter, but that su command is not right. You forgot the -c flag. Also, Solaris (and probably some other systems) want you to quote the command: su amanda -c "amlabel -f amanda.conf DailySet1-000" If that's not the problem, then as Karl said, the error probably means "amlabel" is not a valid binary (program) file for this machine. You say you compiled it on Solaris8i386. Is that also where you are trying to execute it on? For instance, if you take a Solaris program to a Linux machine, it isn't going to work. What do you get if you run this: su amanda -c "ksh -c 'whence amlabel'" Did you get back the path to amlabel you expected? What happens if you run "file" on that path? For instance: # su amanda -c "ksh -c 'whence amlabel'" /usr/local/amanda/sbin/amlabel # file /usr/local/amanda/sbin/amlabel /usr/local/amanda/sbin/amlabel: ELF 32-bit MSB executable SPARC Version 1, dynamically linked, not stripped >Bernd Zimmermann John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: SGIs not backing up
>... You see I am trying to >fix this for a network that I am not familiar with. And the person who >set it up has left. Understood. That wasn't very nice of them :-). >The server is running "Amanda-2.5.0-19990816" >and the SGI clients "Amanda-2.4.2-19990203". As I recall, that's the setup that does not work (2.5 server talking to a pre-2.5 client). And I'm sorry, but I just can't handle digging into it at the moment. It definitely needs fixed, though, since it will be unreasonable to force Amanda users to upgrade everything at once (or clients before server). >I certainly don't want to be on the bleeding edge, but I'm not sure how >much effort will be required to reconfigure the server and all the Sun >clients to the earlier version. ... I'm guessing it would not be too hard. There certainly aren't any major differences in any of the configuration files. I think it would be worth giving it a shot (make a tar copy of your config area before running any of the 2.4.2 stuff, just in case). And if you run into trouble, you know where to ask :-). >> Actually, I suspect the fix is going to require some reasonably >> good coding, C level and protocol debugging experience on your part. > >I'm afraid I can't spare the time to do this. Also understood :-). >Ross John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Failed queue message - Amanda 2.4.2p1 - Investigation results
>In Solaris 8 the raw devices are set to > >crw--- root sys . Actually, in all versions of Solaris (and many other OS's). >ufsdump does a check on the real user id and will not let you at the raw >device if you are not really root even though the link from the /dev >directory is setuid. ... Ummm, that's a bit confusing. It's ufsdump that's setuid, not the /dev directory. But ufsdump only needs that to do remote backups and even then drops permissions as soon as possible. So for all intents and purposes, you can think of it as running as the login and group(s) of whoever started it. >Chmod to allow crw-r--r-- allowed this to work. I >will also add amanda to the sys group and re-chmod to crw-r-. This >should provide a little more security (as these machines are on the >net). Gee, just like the documentation tells you to :-): 2.2. SET UP THE BACKUP CLIENT HOSTS ... B. Set up your raw disk devices so that the dumpuser can read them, and /etc/dumpdates so that the dumpuser can write to it. Normally this is done by making the disk devices readable by (and dumpdates read/writable by) group `operator', and putting the dumpuser into that group. And don't forget to change the group (or however you decide to set things up) on /etc/dumpdates. >David Logan John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Best config for a single large root partition
> 01 Because of dump don't support exclude file lists/patterns, I have to
>use
> gtar, if I want to benefit from this feature?
Correct.
Dump also does not support (in general) anything other than a whole file
system, so that also points you toward using GNU tar.
> 02 If I create an exclude file for the '/home' disklist entry, what about
> the files in '/' ? If I include an entry like '/*' in those exclude
>list,
> '/home' won't be tar'ed too?
I think the idea is that you do *not* define an exclude list in the base
dumptype used for / and /home, and you use that "as is" for /home (i.e.
everything under /home will be processed), but in the disklist entry for /
(or an alternate dumptype) you have an exclusion for /home (etc).
your.host/home root-tar
your.host/var root-tar
your.host/ {
root-tar
exclude list "/path/to/file"
}
> 01 My assumption is, that the amanda exclude "string" (not using the
>list
>option) is passed to the tar parameter that expects the globbing
>file
>PATTERN?
Correct.
> 02 If yes, can I define in the local dumptype an exclude PATTERN for
>the
>appropriate disk entry, so I don't have to create and maintain a
>corresponding exclude list file at the client side? Which will
>mean,
>can I maintain all exclude PATTERNS in the disklist file?
Yes, as long as you only need one pattern, since you can only pass one
--exclude parameter to GNU tar (or at least that's all Amanda can handle).
But in your case, that's not going to work because you want to exlude
both /home and /var from the / dump. Unless you can come up with a
pattern that will do that.
>mike
John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: amandad stand alone????
>Can anyone tell me if it is possible to run amanda in a stand alone config >(i.e. not via inetd or brethren)? Not yet. It's on the list of things to do. >I am trying to set up amanda on a group of external servers. Because >of the limited number of services offered we don't/won't want to run >inetd. ... Huh? Why? I have inetd running on machines with only a couple of services enabled (Amanda being one of them). In the Unix way, use the right tool for the job. Inetd knows all that icky stuff about waiting for connections and starting a process. Why duplicate all that in amandad (and amindexd and amidxtaped)? On the assumption (and it's a big one :-) inetd/xinetd already are doing the right security things, either via tcpd or directly, going to a dedicated amandad opens up yet another place where a hole could be. >What we want to do is be able to start and stop amandad only >during the backup window. ... This can easily be done by swapping the inetd.conf and HUP'ing it. Or you could point inetd.conf at a wrapper around amandad that is time of day (or whatever) sensitive. >-Dustin John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: "no estimate" after 2.4.2p1 upgrade
Ben Jarvis <[EMAIL PROTECTED]> writes: > I recently upgraded from 2.4.1p1 to 2.4.2p1, and ever since I've >had a funny problem. Several of the disks repeatedly say "no estimate" >when amdump runs ... Ben and I worked out this problem offline (thanks, Ben!). Whenever amcleanup is run (which Ben says he did a lot during the upgrade :-), it looks for anything in the holding disk and marks it "force-no-bump" so the same level will be done on the next run. It does this because we don't know for certain the image is complete and it's safer to ask for the same one again. But there was a bug in planner when the level was 0 and this flag is set. During dump analysis, it might decide to do an incremental since it really wasn't time to do another level 0, but earlier code did not ask for a level 1 estimate, which lead to the "no estimate" error. Ben tested the following, and I reproduced the problem and tested it as well. The fix is in the current source tree, but in case you're having this problem, you can either apply this patch: http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/amanda/server-src/planner.c.diff?cvsroot=amanda&only_with_tag=amanda-242-branch&r1=text&tr1=1.76.2.17&r2=text&tr2=1.76.2.18&r2=text&f=u Or you can use amadmin to "unforce-no-bump" the affected clients/disks (although you might want to also "force" them to make sure you have a valid level 0). It's still possible for planner to decide to do an incremental and thus to go down the "no estimate" path, but only if the estimate of the level 0 also failed, in which case you're in deep doo-doo anyway :-). >Ben John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: exclude list format
I understand you figured out the syntax (congratulations! -- it always gives me a headache :-), so the following is just rambling. >The problem is that even though I am telling it to exclude these files AND >use tar it still creates the tarball with those directories which causes >the dump to fail. What is the syntax of the exclude file. ... FYI, I wrote some scripts to help me figure out and test exclude patterns: ftp://gandalf.cc.purdue.edu/pub/amanda/gtartest-exclude >Does anyone >know of a GOOD place for amanda docs. All I can find says use fearture A >but gives no way of how to use feature A. For example the man page talks >about the exclude file but does not cover the format of the exclude file. Have you looked at http://www.backupcentral.com/amanda.html? There is a section in there about exclusion patterns, although I had to fight through figuring them out, so make no guarantees they are 100% correct. As to the exclude syntax specifically, why should Amanda document GNU tar parameters? If you're having trouble (as many, **many**, people do, including myself :-) with a feature of GNU tar, it seems to me that's the place you should go for documentation. On the other hand :-), Amanda is a (the) *heavy* user of this feature, so it would probably be good to write a docs/EXCLUDE file for release. At a minimum, we should at least make it clear that "exclude" maps to "--exclude=PATTERN" and "exclude list" maps to "--exclude-from=FILE" so you could have some hope of finding the right stuff in the GNU tar docs. Any volunteers? (yeah, I thought not :-). >Why does amcheck not find these errors? What errors? Amcheck knows nothing about exclusions (although I think it will check for the file). Are you asking that it run GNU tar and make sure doing it with the exclusions and without generates different results? That would be a bit, shall we say, time consuming. >Andrew Hall John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: amrecover vs. amrestore
>Can someone give me an idea of amrecover vs. amrestore vs. just getting >the tars off the tape manually? I don't understand your question. What kind of idea are you looking for? If you're looking for information on how to use these three methods of getting an image back, they are all covered in: http://www.backupcentral.com/amanda.html >Mark John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: amrecover vs. amrestore
On Thu, 15 Mar 2001, John R. Jackson wrote: > If you're looking for information on how to use these three methods of > getting an image back, they are all covered in: > > http://www.backupcentral.com/amanda.html Just did that. All clear now! (But I'll be back with more questions!) -- http://www.mchang.org/ http://decss.zoy.org/
taper: FATAL attach_buffer: mmap: Invalid argument
I've just installed the amanda 2.4.2p1 server on a Debian Linux 2.2
box.
I keep getting the below error message, although amanda seemed
to have worked. I can't find any docs nor any one else who has
had this problem. What does the error message mean and how do I fix it?
Thanks,
Terry
The next tape Amanda expects to use is: noemix-daily01.
FAILURE AND STRANGE DUMP SUMMARY:
taper: FATAL attach_buffer: mmap: Invalid argument
STATISTICS:
Total Full Daily
Estimate Time (hrs:min)0:00
Run Time (hrs:min) 0:00
Dump Time (hrs:min)0:00 0:00 0:00
Output Size (meg) 0.20.20.0
Original Size (meg) 0.20.20.0
Avg Compressed Size (%) -- -- --
Filesystems Dumped2 2 0
Avg Dump Rate (k/s) 638.2 638.2--
Tape Time (hrs:min)0:00 0:00 0:00
Tape Size (meg) 0.00.00.0
Tape Used (%) 0.00.00.0
Filesystems Taped 0 0 0
Avg Tp Write Rate (k/s) -- -- --
NOTES:
planner: Adding new disk plant:/home/koyama/bin.
DUMP SUMMARY:
DUMPER STATSTAPER STATS
HOSTNAME DISKL ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS
KB/s
-- -
keenan -koyama/bin 0 192192 --0:001403.8 FAILED
plant -koyama/bin 0 32 32 --0:00 148.5 FAILED
(brought to you by Amanda version )
Re: taper: FATAL attach_buffer: mmap: Invalid argument
>I keep getting the below error message, although amanda seemed >to have worked. ... It most definitely did **not** work. The dumps may have happened into the holding disk, but they did not make it to tape. >I can't find any docs nor any one else who has >had this problem. What does the error message mean and how do I fix it? First, "man mmap" and look for EINVAL. Amanda is just reporting to you what the OS told it. Memory mapping is an operating system service that allows processes to share memory. There are two types supported by Amanda, mmap and shmget. My guess is you either do not have these services configured into your system, or they are limited in some way (i.e. the amount of shared memory space is controlled but too small). Beyond that, I can only guess as I don't have experience on your particular OS. Time to crack open the OS documentation or ask on one of those newsgroups/mailing lists. If you do get it worked out, please post the solution back here so it can be in the archives, and/or add it to the FAQ. >Terry John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
