Re: Error msg interpretation
* Ben Elliston [EMAIL PROTECTED] (Tue, Jan 30, 2001 at 10:10:15AM +1000) jrj wrote: Is this 2.4.2? If so, I think you're not the only one who's mentioned this. Sigh. Yes, it is. Interestingly, I think this is the first partition of mine to receive a level 3 dump. Are you using tar to dump ? Level 3 dumps might take a long while to start receiving data when using tar. tar is slow .. It takes currently well over 4 hours for tar to give me an estimate of my disks [rant, rant, fume fume ;) ] more on this in a later post ;) Gerhard, @jasongeo.com == The Acoustic Motorbiker == -- __O Remember that two wrongs never make a right, =`\, But that three do. (=)/(=)
Re: amrecover: cannot connect to host (connection refused)
* Sebastian Frankfurt [EMAIL PROTECTED] (Tue, Jan 30, 2001 at 10:27:41AM +0100) Mack Earnhardt wrote: I was getting connection refused on amrecover for a while. The ..amandahosts file allowed access to root@localhost, but amrecover would use root@servername and fail. My solution was to use the server name explicitly in both .amandahosts and disklist. Hope this help one of your problems. :) Well thanx for the tip, but it doesn't matter, what I put into the .amandahosts because the error I get is: AMRECOVER Version 2.4.1p1. Contacting server on localhost ... amrecover: Unexpected server end of file Did you check the files in /tmp/amanda (on server and client) ? Kind regards, -- Gerhard den Hollander Phone +31-10.280.1515 Technical Support 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: amrecover: cannot connect to host (connection refused)
Gerhard den Hollander wrote: * Sebastian Frankfurt [EMAIL PROTECTED] (Tue, Jan 30, 2001 at 10:27:41AM +0100) Mack Earnhardt wrote: I was getting connection refused on amrecover for a while. The ..amandahosts file allowed access to root@localhost, but amrecover would use root@servername and fail. My solution was to use the server name explicitly in both .amandahosts and disklist. Hope this help one of your problems. :) Well thanx for the tip, but it doesn't matter, what I put into the .amandahosts because the error I get is: AMRECOVER Version 2.4.1p1. Contacting server on localhost ... amrecover: Unexpected server end of file Did you check the files in /tmp/amanda (on server and client) ? I have had a look at them don't see anything wrong :-( amrecover: amrecover: debug 1 pid 1554 ruid 0 euid 0 start time Tue Jan 30 10:28:33 2001 Error connecting to server amrecover: pid 1554 finish time Tue Jan 30 10:28:33 2001 I will attach all /tmp/amanda files. thanx, Sebastian PS: rsh commands works properly for root and operator amandad: debug 1 pid 1495 ruid 11 euid 11 start time Tue Jan 30 10:25:31 2001 amandad: version 2.4.1p1 amandad: build: VERSION="Amanda-2.4.1p1" amandad:BUILT_DATE="Mon Aug 21 16:31:37 EDT 2000" amandad:BUILT_MACH="Linux porky.devel.redhat.com 2.2.5-22smp #1 SMP Wed Jun 2 09:11:51 EDT 1999 i686 unknown" amandad:CC="gcc" amandad: paths: bindir="/usr/bin" sbindir="/usr/sbin" amandad:libexecdir="/usr/lib/amanda" mandir="/usr/share/man" amandad:CONFIG_DIR="/etc/amanda" DEV_PREFIX="/dev/" amandad:RDEV_PREFIX="/dev/r" DUMP="/sbin/dump" amandad:RESTORE="/sbin/restore" SAMBA_CLIENT="/usr/bin/smbclient" amandad:GNUTAR="/bin/tar" COMPRESS_PATH="/usr/bin/gzip" amandad:UNCOMPRESS_PATH="/usr/bin/gzip" MAILER="/usr/bin/Mail" amandad:listed_incr_dir="/var/lib/amanda/gnutar-lists" amandad: defs: DEFAULT_SERVER="localhost" DEFAULT_CONFIG="DailySet1" amandad:DEFAULT_TAPE_SERVER="localhost" amandad:DEFAULT_TAPE_DEVICE="/dev/null" HAVE_MMAP HAVE_SYSVSHM amandad:LOCKING=POSIX_FCNTL SETPGRP_VOID DEBUG_CODE BSD_SECURITY amandad:USE_AMANDAHOSTS CLIENT_LOGIN="operator" FORCE_USERID amandad:HAVE_GZIP COMPRESS_SUFFIX=".gz" COMPRESS_FAST_OPT="--fast" amandad:COMPRESS_BEST_OPT="--best" UNCOMPRESS_OPT="-dc" got packet: Amanda 2.4 REQ HANDLE 000-90D50508 SEQ 980846701 SECURITY USER operator SERVICE sendbackup OPTIONS hostname=localhost; GNUTAR /etc 1 2001:1:27:2:25:7 OPTIONS |;bsd-auth;compress-fast;index;exclude-list=/usr/local/lib/amanda/exclude.gtar; sending ack: Amanda 2.4 ACK HANDLE 000-90D50508 SEQ 980846701 bsd security: remote host localhost user operator local user operator amandahosts security check passed amandad: running service "/usr/lib/amanda/sendbackup" amandad: sending REP packet: Amanda 2.4 REP HANDLE 000-90D50508 SEQ 980846701 CONNECT DATA 1297 MESG 1298 INDEX 1299 OPTIONS ;compress-fast;bsd-auth;index;exclude-list=/usr/local/lib/amanda/exclude.gtar; amandad: got ack: Amanda 2.4 ACK HANDLE 000-90D50508 SEQ 980846701 amandad: pid 1495 finish time Tue Jan 30 10:25:31 2001 amrecover: debug 1 pid 1639 ruid 0 euid 0 start time Tue Jan 30 10:59:36 2001 Error connecting to server amrecover: pid 1639 finish time Tue Jan 30 10:59:36 2001 amtrmidx: debug 1 pid 1541 ruid 11 euid 11 start time Tue Jan 30 10:25:32 2001 /usr/lib/amanda/amtrmidx: version 2.4.1p1 Keeping 26 index files localhost /etc amtrmidx: pid 1541 finish time Tue Jan 30 10:25:40 2001 killpgrp: debug 1 pid 2007 ruid 11 euid 0 start time Sat Jan 27 03:16:49 2001 /usr/lib/amanda/killpgrp: version 2.4.1p1 sending SIGTERM to process group 2007 child process exited with status 3 runtar: debug 1 pid 1500 ruid 11 euid 0 start time Tue Jan 30 10:25:31 2001 gtar: version 2.4.1p1 running: /bin/tar: gtar --create --directory /etc --listed-incremental /var/lib/amanda/gnutar-lists/localhost_etc_1.new --sparse --one-file-system --ignore-failed-read --totals --file - . selfcheck: debug 1 pid 1453 ruid 11 euid 11 start time Tue Jan 30 10:20:01 2001 /usr/lib/amanda/selfcheck: version 2.4.1p1 checking disk /etc: device /etc: OK selfcheck: pid 1453 finish time Tue Jan 30 10:20:01 2001 sendbackup: debug 1 pid 1496 ruid 11 euid 11 start time Tue Jan 30 10:25:31 2001 /usr/lib/amanda/sendbackup: got input request: GNUTAR /etc 1 2001:1:27:2:25:7 OPTIONS |;bsd-auth;compress-fast;index;exclude-list=/usr/local/lib/amanda/exclude.gtar; parsed request as: program `GNUTAR' disk `/etc' lev 1 since 2001:1:27:2:25:7 opt `|;bsd-auth;compress-fast;index;exclude-list=/usr/local/lib/amanda/exclude.gtar;' sendbackup: exclude list file "/usr/local/lib/amanda/exclude.gtar" does not exist, ignoring waiting for connect on 1297, then 1298, then 1299 got all connections sendbackup: spawning "/usr/bin/gzip" in pipeline
Getting estimates ..
getting estimates with gnuatr takes forever and then some. Amanda also comes with a little tool called calcsizes that 1) seems to be fatser than gnutar 2) can do mutlipel estimates in one run. Currently getting an estimate on the big disk takes 3 -4 hours and that's mainly because for every levbel, a new gnutar is being run. So to get estimates for a level 0, 1 and 2 run I see amanda doing get gnutar estimate for level 0 (this takes over an hour) when this has finished get gnutar estimate for level 1 (this takes over an hour) when this has finished get gnutar estimate for level 2 (this takes over an hour) total time, close to 4 hours. using calcsize, I should be able to do something like calcsize GNUTAR lto /whopper/home 0 0 1 980801273 2 980851273 and get the estimates for level 0 1 and 2 all at once. So my question is, How do I tell amanda to use calcsize ? I didn't find anything in docs/* man/* or config/* %grep calcsize docs/* man/* config/* docs/WHATS.NEW:"calcsize" and "runtar" therefore must be installed setuid root. I've made If this is not possible, can this please be taken as a serious wishlist for Amanda 2.5 ? Also, if this is not possible, does anyone have any other suggestions on getting a speedup on the estimates ? The sendsize.c file menations a maxdumps= variable .. How can this be set ? Kind regards, -- Gerhard den Hollander Phone +31-10.280.1515 Technical Support 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.
amdump and eject
Hey folks, Is this safe to put in the Amanda crontab? /usr/sbin/amdump Daily; /usr/bin/mt -f /dev/nst0 eject I'm just wondering if "amdump" spawns other jobs that need to be executed before quitting and then ejecting the tape. Cheers, Luc. PS: Amanda is really great. I've only one complaint though. I can't seem to get estimates from Solaris 2.7. "sensize" takes forever with "tar". I eventually have to kill it. I just bypass the problem and use "ufsdump". Is this one of the bugs that is fixed in 2.4.2p1? -- Luc Lalonde, Responsable du reseau GIREF Telephone: (418) 656-2131 poste 6623 Courriel: [EMAIL PROTECTED] begin:vcard n:Lalonde;Luc x-mozilla-html:FALSE org:Universite Laval;GIREF adr:;; version:2.1 email;internet:[EMAIL PROTECTED] title:Administateur de reseau x-mozilla-cpt:;0 fn:Luc Lalonde end:vcard
exclude is not working .. :?
Hi all
I have a problem ...
I use this dumptype to dump my / filesystem (I use the /dev/sda1 in the
disklist file) :
define dumptype comp-root-tar {
root-tar
comment "Root partitions with compression"
compress client fast
exclude list "/usr/local/etc/amanda/Linux/root-exclude"
index yes
}
The exclude list had
/var/spool
and it did report those problems when backing up ... :
? gtar: Cannot add file ./var/spool/postfix/active/5/8/58FEE3D040: No such file or
directory
? gtar: Cannot add file ./var/spool/postfix/active/D/9/D90CB3D024: No such file or
directory
(obviously by the time it tried to dump the files, those were not there
already)
so, I see the "." prepending the path, and added another path to the exclude
list :
/var/spool
./var/spool
but the error persists :/
could anyone let me know why ? :?
--
Javi Polo - [EMAIL PROTECTED] - navo - DrSlump
Proud member of the Panda Gey Community (powered by linux)
http://javipolo.ivworlds.org/ - Fidonet 2:347/1.1
Re: amdump and eject
* Luc Lalonde [EMAIL PROTECTED] (Tue, Jan 30, 2001 at 08:55:29AM -0500) Hey folks, Is this safe to put in the Amanda crontab? /usr/sbin/amdump Daily; /usr/bin/mt -f /dev/nst0 eject I have it in mine. I'm just wondering if "amdump" spawns other jobs that need to be executed before quitting and then ejecting the Not that I have noticed. PS: Amanda is really great. I've only one complaint though. I can't seem to get estimates from Solaris 2.7. "sensize" takes forever with "tar". tar takes forever. I've hacked sendsize to use calcsize instead (which Im testing now on my linux box , where I can afford to screw up the backup process once or twice). using the hack below the estimate fase takes 16 minutes iso 45 - 60 minutes. It's currently backing up , Im curious to see how it pans out. on my big disk gnutar takes 3 times 75 minutes to ghet level 0, 1 and 2 estimates of a 110G dir-tree calcsize takes 35 minutes (and does all 3 levels in one go) Note, the hack below is not fully functional yet, and will *always* use calcsize, even if you're using ufsdump !!. In sendsize.c, change the bit from line 370 onward into /* Now in the child process */ #ifndef USE_GENERIC_CALCSIZE if(strcmp(est-program, "DUMP") == 0) dump_calc_estimates(est); else # ifdef SAMBA_CLIENT if (strcmp(est-program, "GNUTAR") == 0 est-amname[0] == '/' est-amname[1] == '/') smbtar_calc_estimates(est); else # endif # ifdef GNUTAR if (strcmp(est-program, "GNUTAR") == 0) gnutar_calc_estimates(est); else # endif #endif generic_calc_estimates(est); if (maxdumps 1) exit(0); } and #define USE_GENERIC_CALCSIZE (e.g. in config/config.h) I eventually have to kill it. I just bypass the problem and use "ufsdump". Is this one of the bugs that is fixed in 2.4.2p1? No. Gerhard, @jasongeo.com == The Acoustic Motorbiker == -- __O "Bipedalism--an unrecognized disease affecting over 99% of the =`\, population. Symptoms include lack of traffic sense, slow rate of (=)/(=) travel, and the classic, easily recognized behavior known as cycling."
Re: exclude is not working .. :?
Hi Javi, I've had 2 problems with this. First of all, try to check what file amanda thinks the gtar exclude files are in--I found out that it was really looking in /usr/local/lib/amanda/exclude.gtar even though it was supposed to be /usr/local/var/amanda/exclude.gtar. I just kept the file in /usr/local/var/amanda and made a symbolic link from /usr/local/lib/amanda so the file would be accessible from both places just in case. Next, check the gtar docs in "info" to see how the exclude paths work. I found they don't work at all like I expected, and I haven't got them to work with file paths, just regexp's. I wouldn't necessarily exclude all of /var/spool unless you don't want to back up all of your printer configurations for example. I don't care for GNU info, so I use emacs then go to "Help-Manuals-Browse Manuals with Info" and I find that's a little easier to navigate (that's on a Linux or IRIX system, don't know if emacs is set up that way on your OS). Javi Polo wrote: so, I see the "." prepending the path, and added another path to the exclude list : /var/spool ./var/spool but the error persists :/ could anyone let me know why ? :? -- "Jonathan F. Dill" ([EMAIL PROTECTED])
Help with error message
Over the weekend, we received this error message from Amanda: FAILURE AND STRANGE DUMP SUMMARY: amanda sd0e lev 1 FAILED [nak error:unexpected ack packet] amanda is the name of the backup server. I checked what I knew to check and could not find anything wrong. The backup last night ran fine. The error does not seem critical, but I would like to know what happened. Can some kind soul explain what this means? We're running amanda 2.4.1p1 on a PC running NetBSD. Thanks! Andrew Robinson * Andrew W. Robinson | Voice: +1 (504)-889-2784 * * Computerized Processes Unlimited, Inc. | FAX:+1 (504)-889-2799 * * 4200 S. I-10 Service Rd., Suite 205| E-Mail: [EMAIL PROTECTED] * * Metairie, LA 70001 | WWW: http://www.cpu.com * * "Consulting System Integrators" *
Calcsize vs gnutar to do estimates
I just hacked sendsize.c to use calcsize iso gnutar to estimate dumpsizes.
The speed difference is enorm.
Whereas gnutar took well over 4 hours to get the estimates using calcsize
gives results within 45 minutes.
Now my question is:
1) Why is gnutar used in stead of calcsize ?
2) Are there any dangers in running calcsize ?
I gather from readiong the source, the changelog, and some searching on the
egroups list that calcsize should give the same numbers as gnutar .
At any rate,
Attached below my (small) hack to sendsize.c
(basically I moved the #ifndef USE_GENERIC_CALCSIZE 3 lines lower).
Im not sure if this is the intedned behaviour
(as it is now, it still uses DUMP to get dump estimates, and calcsize for
everything else. I noticed dump on SGI returning within seconds with a
correct estimate, and calcsize isn;t going to beat that ;) )
May I suggest the following changes for Amanda 2.5 (or whatever :) )
1) change USE_GENERIC_CALCSIZE into USE_CALCSIZE_FOR_GNUTAR
(or something like that ;) )
2) add an option to ./configure to set this as a #define in config.h
(heck, I'd happily do so myself, but I do not know what the procedure is ).
void calc_estimates(est)
disk_estimates_t *est;
{
dbprintf(("calculating for amname '%s', dirname '%s'\n", est-amname,
est-dirname));
if (maxdumps 1) {
while(dumpsrunning = maxdumps) {
wait(NULL);
--dumpsrunning;
}
++dumpsrunning;
switch(fork()) {
case 0:
break;
case -1:
error("calc_estimates: fork returned: %s", strerror(errno));
default:
return;
}
}
/* Now in the child process */
if(strcmp(est-program, "DUMP") == 0)
dump_calc_estimates(est);
else
#ifndef USE_GENERIC_CALCSIZE
# ifdef SAMBA_CLIENT
if (strcmp(est-program, "GNUTAR") == 0
est-amname[0] == '/' est-amname[1] == '/')
smbtar_calc_estimates(est);
else
# endif
# ifdef GNUTAR
if (strcmp(est-program, "GNUTAR") == 0)
gnutar_calc_estimates(est);
else
# endif
#endif
generic_calc_estimates(est);
if (maxdumps 1)
exit(0);
}
gerhard, @jasongeo.com == The Acoustic Motorbiker ==
--
__O "Hello. Welcome to Hell. I am, of course, the Devil. Now, we'll start be
=`\, separating you all into lots. First, the Germans. You knew that was
(=)/(=) coming, didn't you? Over here. Next, the French. You can go stand by
the Germans. Atheists over here. I bet you feel bloody stupid right now
eh? Christians, over here. What? Oh, I'm sorry, but it turns out the
Jews were right...
hostname
Hello everybody: i configured amanda to run as a server in a machine whith an oldhostname Now i changed the hostname to newhostname when i do amrecover it looks for oldhostname. How can i tell amanda that i have changed the name of the server where it is running?
Followup: Client stopped responding, Request timed out
I sent the message below yesterday. After having read some other seemingly unrelated messages, I realized that I did change one of the disks on the client "client.foo.com" from using dump to tar (see dumptypes below). It was after this change that the entire client started coming back as "Request timed out". I got this from reading the other messages saying that backups with tar took an incredibly long time to generate estimates. I want to use tar so that I can exclude some areas of that disk. Is it enough to just increase the etimeout for that client? Am I on the right track? All the logfiles are still included below from the original message. Oscar I've been running Amanda 2.4.2 for several months now. I have a client, client.foo.com that has stopped responding and here's what I get from the Amanda report: FAILURE AND STRANGE DUMP SUMMARY client.foo. /disk1 lev 0 FAILED [Request to client.foo.com timed out.] client.foo. /var lev 0 FAILED [Request to client.foo.com timed out.] client.foo. /usr lev 0 FAILED [Request to client.foo.com timed out.] client.foo. /home lev 0 FAILED [Request to client.foo.com timed out.] client.foo. / lev 0 FAILED [Request to client.foo.com timed out.] Amanda is launched from inetd and I did verify that inetd was running. Both client and amanda server are connected to switched 100 Mbps ports so speed isn't necessarily an issue. I also increased the the timeouts (although ctimeout I just raised from 30 to 60): etimeout1200 dtimeout1800 ctimeout60 Here is the debug files for client.foo.com: amandad.debug amandad: debug 1 pid 10696 ruid 520 euid 520 start time Sun Jan 28 22:30:49 2001 amandad: version 2.4.2 amandad: build: VERSION="Amanda-2.4.2" amandad: BUILT_DATE="Thu Jul 6 13:33:02 CDT 2000" amandad: BUILT_MACH="Linux client.foo.com 2.2.5-22 #1 Wed Jun 2 09:17:03 EDT 1999 i686 unknown" amandad: CC="gcc" amandad: paths: bindir="/usr/local/amanda/bin" amandad: sbindir="/usr/local/amanda/sbin" amandad: libexecdir="/usr/local/amanda/libexec" amandad: mandir="/usr/local/man" AMANDA_TMPDIR="/tmp/amanda" amandad: AMANDA_DBGDIR="/tmp/amanda" amandad: CONFIG_DIR="/usr/local/amanda/etc" DEV_PREFIX="/dev/" amandad: RDEV_PREFIX="/dev/" DUMP="/sbin/dump" amandad: RESTORE="/sbin/restore" SAMBA_CLIENT="/usr/bin/smbclient" amandad: GNUTAR="/usr/local/bin/tar" COMPRESS_PATH="/usr/bin/gzip" amandad: UNCOMPRESS_PATH="/usr/bin/gzip" MAILER="/usr/bin/Mail" amandad: listed_incr_dir="/usr/local/amanda/var/gnutar-lists" amandad: defs: DEFAULT_SERVER="amanda.foo.com" amandad: DEFAULT_CONFIG="daily" amandad: DEFAULT_TAPE_SERVER="amanda.foo.com" amandad: DEFAULT_TAPE_DEVICE="/dev/nrsa0" HAVE_MMAP HAVE_SYSVSHM amandad: LOCKING=POSIX_FCNTL SETPGRP_VOID DEBUG_CODE BSD_SECURITY amandad: USE_AMANDAHOSTS CLIENT_LOGIN="amanda" FORCE_USERID HAVE_GZIP amandad: COMPRESS_SUFFIX=".gz" COMPRESS_FAST_OPT="--fast" amandad: COMPRESS_BEST_OPT="--best" UNCOMPRESS_OPT="-dc" got packet: Amanda 2.4 REQ HANDLE 004-80310808 SEQ 980742604 SECURITY USER amanda SERVICE sendsize OPTIONS maxdumps=2;hostname=client.foo.com; DUMP /drive1 0 1970:1:1:0:0:0 -1 DUMP /drive1 1 2001:1:10:6:20:15 -1 DUMP /var 0 1970:1:1:0:0:0 -1 DUMP /var 1 2001:1:17:2:42:18 -1 DUMP /usr 0 1970:1:1:0:0:0 -1 DUMP /usr 1 2001:1:18:2:25:44 -1 DUMP /home 0 1970:1:1:0:0:0 -1 DUMP /home 1 2001:1:21:5:26:24 -1 GNUTAR / 0 1970:1:1:0:0:0 -1 exclude-list=/usr/local/amanda/exclude GNUTAR / 1 2001:1:20:5:28:22 -1 exclude-list=/usr/local/amanda/exclude sending ack: Amanda 2.4 ACK HANDLE 004-80310808 SEQ 980742604 bsd security: remote host amanda.foo.com user amanda local user amanda amandahosts security check passedamandad: running service "/usr/local/amanda/libexec/sendsize" selfcheck.debug selfcheck: debug 1 pid 16214 ruid 520 euid 520 start time Sun Jan 28 19:00:47 2001 /usr/local/amanda/libexec/selfcheck: version 2.4.2 checking disk /disk1: device /dev/sdb1: OK checking disk /var: device /dev/sda6: OK checking disk /usr: device /dev/sda5: OK checking disk /home: device /dev/sda8: OK checking disk /: device /: OK selfcheck: pid 16214 finish time Sun Jan 28 19:00:47 2001 runtar.debug runtar: debug 1 pid 10702 ruid 520 euid 0 start time Sun Jan 28 22:30:50 2001 /usr/local/bin/tar: version 2.4.2 running: /usr/local/bin/tar: /usr/local/bin/tar --create --directory / --listed-incremental /usr/local/amanda/var/gnutar-lists/client.foo.com__0.new --sparse --one-file-system --ignore-failed-read --totals --file /dev/null --exclude-from /usr/local/amanda/exclude . sendsize.debug sendsize: debug 1 pid 10697 ruid 520 euid 520 start time Sun Jan 28 22:30:50 2001 /usr/local/amanda/libexec/sendsize: version 2.4.2 amandates botch: /dev/sda1 lev 0: new dumpdate 959404817 old 979968536 amandates botch: /dev/sda5 lev 0: new dumpdate 959405582 old 979784774 amandates botch: /dev/sda6 lev 0: new dumpdate 959405079 old 979699366 amandates
Re: exclude is not working .. :?
so, I see the "." prepending the path, and added another path to the exclude list : /var/spool ./var/spool but the error persists :/ could anyone let me know why ? :? Jonathan had a really good idea about making sure GNU tar is even using the exclusions file you think it is. Take a look at the /tmp/amanda/sendbackup*debug file for the arguments passed to GNU tar. Even with the right file :-), I always have a terrible time figuring out tar exclusions. They never seem to work the way I expect/want them too. You might take a look at: ftp://gandalf.cc.purdue.edu/pub/amanda/gtartest-exclude It's a little ksh script that lets you set up test cases and find out what GNU tar will do with them. Javi Polo John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: hostname
... when i do amrecover it looks for oldhostname. How can i tell amanda that i have changed the name of the server where it is running? Do a "man amrecover" and look for the -s and -t flags. If you don't want to use those flags and want it to pick the new name by default, you'll have to "make distclean" and rerun configure, etc, to rebuild amrecover. John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: amanda 2.4.2 and Changer problems
Tue Jan 30 0:04:03 2001: enter: getTapeParams Tue Jan 30 0:04:03 2001: running: /bin/chio -f /dev/null params I don't know much about chio, but passing /dev/null to it seems like a bad idea. That comes from "changerdev" in amanda.conf. Are you sure that parameter didn't get dinged in the upgrade or testing? Clem Kumah John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before?
Title: Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before? Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before? Running an Exabyte 8500 series tape drive. If so please!!! Tell me how you did it. Christopher A. Los Angeles, Ca.
dumps (barely) too big
I'm using amanda-2.4.2, and I've got one filesystem that's (after compression) 500MB too big for my tape. :( I'd rather not use gtar, due to previous bad experiences with it. Does 2.5 out of CVS support splitting dumps across multiple tape volumes? Chris -- - [EMAIL PROTECTED] Chris JonesSRI International, Inc.
Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before?
Title: Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before? Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before? Running an Exabyte 8500 series tape drive. If so please!!! Tell me how you did it. Christopher A. Los Angeles, Ca.
Planner - balance suggestions?
Hey all, I've got a puzzler here, any suggestions on how to handle this would be great (except using gnutar). I'm using amanda-2.4.2 on Solaris 2.6, works great. However, my file systems are such that I have three file systems that are an order of magnitude larger than the rest(18GB vs 2 GB). The problem comes to the balance calculation. The larger file systems never get promoted, so they are left to be dumped until they are actually due. That isn't a problem until 2 biggies are due on the same day, resulting in backup running until noon. :) I had thought no big deal, I'll set the dumpcycle on the biggies to a shortened interval and planner will figure it out. But, the balance size is calculated based on the estimated runspercycle and total size of level 0 dumps, so that adjustment didn't change anything other than increase the frequency of "overtime" backups. No help. As always, in the act of describing the problem, I think I see the solution. I should change my dumpcycle to shorter value and override that on my smaller filesystems to a longer value. That should bump up the balance calculation. Make sense? Later, Bill Carlson -- Systems Programmer[EMAIL PROTECTED]| Opinions are mine, Virtual Hospital http://www.vh.org/| not my employer's. University of Iowa Hospitals and Clinics|
Re: tapelists permissions being changed...
and sure enough, if I check out the permissions: -rw---1 root disk 104 Jan 30 01:05 tapelist I can change the permissions, but they'll be reset after the next dump. Which program writes to this file? ... Taper (amdump), amadmin and amlabel. I'm assuming that I have a broken sticky bit somewhere... I assume you mean setuid (s), not sticky (t). but I think I've been hit with the dumb stick this week - can't find it for the life of me (I checked through the list of setuids that you sent me last time John). It's unlikely you're running amlabel, so that pretty much leave amadmin or taper (from within amdump). So, do you call amadmin as any part of your dump sequence? Do you call it as the Amanda user? Are you positive you are running amdump as the Amanda user? Who owns, for instance, the amdump.1 file? If it's root, too, then amdump is being run by root. Taper should most definitly **not** be setuid-root, so that could be another problem. Chris John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Calcsize vs gnutar to do estimates
Now my question is: 1) Why is gnutar used in stead of calcsize ? For one thing, it knows about exclusion lists. Whether you can actually write one that does what you want is a separate question :-). My guess is the original sequence of events went something like this: * Using tar to /dev/null to gather the estimates used to be a bad idea because it actually read and wrote the data. * Someone wrote calcsize to do something close enough to what tar did to make planner happy. * Later on, tar was changed to detect output was /dev/null and go into a "fast" mode where it did everything but the actual I/O. * The Amanda folks at the time thought this was a better plan (it's usually a good idea to use the same tool for the same job) so changed the code to use tar and sidelined calcsize. Another reason to use GNU tar is the DUMPER-API, if anyone ever implements it, probably won't get along with this change as it looks at these steps from a higher level. I doubt (although I didn't check) they were planning on the estimate hook getting called only once for all requested levels. That might just need a (minor) change to the API, though. And now would certainly be the time to do such a thing. 2) Are there any dangers in running calcsize ? It hasn't been used in years, so who knows what it does any more. It may or may not always give reasonable results (or even build) on all OS's. There was serious consideration of removing it a while back and I think the only reason it's still around is nobody got up enough energy to zap it. Don't take this wrong. I'm all in favor of anything that makes Amanda better. But it sounds to me like someone should be profiling GNU tar and finding out what its problem is instead of hacking around it in Amanda. gerhard John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Client stopped responding, Request timed out
Yes, this was all that was in the file and I now see that it was incomplete. I didn't see the process running so it was either killed or it crashed. I'll stay with tar and the increased timeout for tonight and see if the process does complete. Oscar At 05:20 PM 1/30/01 -0500, John R. Jackson, you wrote: client.foo. /var lev 0 FAILED [Request to client.foo.com timed out.] ... Here is the debug files for client.foo.com: ... sendsize: debug 1 pid 10697 ruid 520 euid 520 start time Sun Jan 28 22:30:50 2001 ... DUMP: mapping (Pass I) [regular files] Total bytes written: 47718400 (45MB, 2.8MB/s) . Was this really all that was in the file? If so, it either crashed or is still running, either of which would explain the timeout. Is it enough to just increase the etimeout for that client? Maybe. If it ever completes, you can look at the first and last line of sendsize*debug for the timestamps and see how long you have to set etimeout to. Note that you cannot set the timeout just for this client. That is a global (all clients) value, not that it really matters. John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: amrecover: cannot connect to host (connection refused)
This is what the amrecover sends to stdout: AMRECOVER Version 2.4.1p1. Contacting server on localhost ... amrecover: Unexpected server end of file And this is what I found in /var/log/messages: ... Jan 29 07:52:07 myhost xinetd[1318]: warning: can't get client address: Invalid argument Stan Brown worked on this almost a year ago and came up with the following patch, which I lost in my inbox (sigh). Please give it a try and let me know if it works so I can (finally) get it into a release. Jan 29 07:52:07 myhost xinetd[1318]: file descriptor of service amandaidx has been closed ... This happens only the very first time, if I try to connect via amrecover the next time, nothing happens in the syslog, but the amrecover client barfs: AMRECOVER Version 2.4.1p1. Contacting server on localhost ... amrecover: Error connecting to server: Connection refused That's because xinetd was **very** unhappy :-) and shut down the amandaidx service (see the message it logged). So there isn't anyone listening any more. I don't know how xinetd does it, but other versions of inetd will reenable a service after (e.g.) 10 minutes, or if you HUP them. Sebastian John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED] amrecover.diff
Re: Getting estimates ..
Also, if this is not possible, does anyone have any other suggestions on getting a speedup on the estimates ? Fix GNU tar? :-) The sendsize.c file menations a maxdumps= variable .. How can this be set ? With "maxdumps" in amanda.conf. It will cause more than one estimate (on separate disks) to be done at the same time, just like during dumping. Note that it will not run multiple estimate levels of the same disk at the same time (I don't think) since that would normally beat on the disk too hard. Gerhard den Hollander John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Quick question about STCTL
Title: Quick question about STCTL I'm trying to get stctl working so that I can move on to the Amanda stages, Can anyone help me with this that has probably had troubles in my same area? I'm situation is below: I have stctl installed and configured (I think properly) on my Sun E250 running Solaris 7 in 64 bit mode (I am still using the 32 bit driver nonetheless). I have an Exabyte 10 H changer assembly with an Exabyte 8500 series drive. When I use the command 'stc status' I get this output: numdrives=1(0) numtransports=1(e) numloadbays=0() numslots=10(m,e,e,e,e,e,e,e,e,e) And when I type the command: 'stc load 3 1' (assuming I'm saying load tape 3 into drive 1) I get this output: stc: Slot 3 reported empty This isn't true, I have tapes loaded in every slot on the changer. Can you think of anything that I might be doing wrong? Assuming I have configured the /usr/kernel/drv/stctl.conf correctly and the major and minor updates are correct is there a way to cause stc to learn that the slots are full? Or is there a way for stc to check to see if the bays have tapes in them so that it doesn't report 'stc: Slot # empty' ? Thanks for all of your help, I have been on this for months with no success Christopher A. Starbase Corp. Los Angeles, Ca.
Re: why | ufsrestore?
Does increasing tapebufs improve the speed in which amanda dump to disk ? No. I thought tapebufs was only for dumping to tape ? Yes. Gerhard John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: updated mtx script?
On Sat, Jan 27, 2001 at 01:21:43AM -0800, Joe Rhett wrote: My diff doesn't understand "-u" as an option. I've attached a normal diff and a context diff between the original chg-zd-mtx.sh.in and my update. A context diff is fine. It works perfectly as a drop-in replacement. It also - * fixes numerous bugs * skips empty slots * works with mtx.1.2.x * supports tape drives which autoclean * supports tape drives which require 'mt offline' before 'mtx unload' Thanks Joe, The patch looks OK, but I can't test it myself. I would appreciate if someone else could test it and report success or failure. Jean-Louis -- Jean-Louis Martineau email: [EMAIL PROTECTED] Departement IRO, Universite de Montreal C.P. 6128, Succ. CENTRE-VILLETel: (514) 343-6111 ext. 3529 Montreal, Canada, H3C 3J7Fax: (514) 343-5834
Re: amreport broken between 2.4.1p1 and 2.4.2
On Sun, Jan 28, 2001 at 07:55:57PM -0500, John R. Jackson wrote: The email output of amreport seems to have been broken somewhere between 2.4.1p1 and 2.4.2. What you call "broken", others call a new "feature" :-). Look for "columnspec" in "man amanda". FYI, here's what I use: columnspec "OrigKB=1:8,OutKB=1:8,DumpRate=0:7,TapeRate=0:7" It would also make more sense (to me, at least) for the output figures to be represented in Mb and Mb/sec rather than Kb and Kb/sec. This would probably help the email readability. Patches are welcome. Probably as some extension to columnspec (maybe a multiplication factor and another field for the column heading text?). Henning P. Schmiedehausen posted a patch on amanda-hackers on november 2000 under the subject "ADR50 Support and reporter Patches for Amanda 2.4.2b2". You could start from his work. Jean-Louis -- Jean-Louis Martineau email: [EMAIL PROTECTED] Departement IRO, Universite de Montreal C.P. 6128, Succ. CENTRE-VILLETel: (514) 343-6111 ext. 3529 Montreal, Canada, H3C 3J7Fax: (514) 343-5834
Re: Has anyone EVER gotten Amanda to work with an Exabyte 10 H change r before?
Yes, but we'd need to know that OS, hardware, etc you are running before we can help you in any useful fashion. If I told you "Yes, running my private distribution of Linux" it wouldn't help you much ;-) The actual answer is 'yes, running Solaris 8 with the sgen driver and mtx' but again this may or may not be helpful. On Tue, Jan 30, 2001 at 11:58:29AM -0800, Adams, Christopher wrote: Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before? Running an Exabyte 8500 series tape drive. If so please!!! Tell me how you did it. Christopher A. Los Angeles, Ca. -- Joe Rhett Chief Technology Officer [EMAIL PROTECTED] ISite Services, Inc. PGP keys and contact information: http://www.noc.isite.net/Staff/
RE: Has anyone EVER gotten Amanda to work with an Exabyte 10 H change r before?
Title: RE: Has anyone EVER gotten Amanda to work with an Exabyte 10 H change r before? I'm running Solaris 7 in 64 bit mode with the stctl driver, but the stctl driver keeps reporting all the bays empty. The connection type from Sun Box to Exabyte changer is SCSI II. Let me know if I can supply you with anymore needed information. Sorry about being vague before, its just that last time I asked this mailing list about this particular problem I wrote about 3 pages of details and got not a single bite. I was kind of testing to see if I got a response this time. Glad you responded, I'll let you know everything I can. -Original Message- From: Joe Rhett [mailto:[EMAIL PROTECTED]] Sent: Tuesday, January 30, 2001 4:54 PM To: Adams, Christopher Cc: '[EMAIL PROTECTED]' Subject: Re: Has anyone EVER gotten Amanda to work with an Exabyte 10 H change r before? Yes, but we'd need to know that OS, hardware, etc you are running before we can help you in any useful fashion. If I told you Yes, running my private distribution of Linux it wouldn't help you much ;-) The actual answer is 'yes, running Solaris 8 with the sgen driver and mtx' but again this may or may not be helpful. On Tue, Jan 30, 2001 at 11:58:29AM -0800, Adams, Christopher wrote: Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before? Running an Exabyte 8500 series tape drive. If so please!!! Tell me how you did it. Christopher A. Los Angeles, Ca. -- Joe Rhett Chief Technology Officer [EMAIL PROTECTED] ISite Services, Inc. PGP keys and contact information: http://www.noc.isite.net/Staff/
Re: why | ufsrestore?
How does one configure the blocksize? Marc was (I assume) referring to old, crufty, icky, by hand scripts he used in the far past, not wonderful, shiny Amanda :-). I fully intend to ignore the fact that they were faster :-). I think what he meant was he changed the 'b' flag value on the dump, which increases the size of the write() call (and possibly some network ioctl sizes). I just ran a few quick tests on a system here with a 500 MByte image: sendsize - dumper - holding disk 1994.1 sendsize -t (client only to /dev/null) 2104 simulated sendsize with named pipes 2103 ufsdump to /dev/null (no ufsrestore) 2118 Without going into detail about the various tests, they are all so close on my system as to make essentially no difference (~ 1%). What about the blocksize used on the tape? perhaps that can be tuned, too... Eventually, but this will have a lot of ripple effects (like how you do restores without Amanda). And it doesn't affect the problem under discussion in this thread, which is why the dump to disk (actually, to/through the network) is slower than would seem reasonable. g. John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
Re: Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before?
I believe mine was either a 10h or a 10i. Anyway, I had to use Fermilab's Juke software. Depends on the OS you are running as well. This was under Digital Unix 4.0x. It worked out pretty dang good. I suspect you will have more choices of robot software if you are using Linux or Solaris. -Dan PS - Don't ask me for my scriptsthat was 3-4 years ago...before I archived my stuff :) On Tue, 30 Jan 2001, Adams, Christopher wrote: Has anyone EVER gotten Amanda to work with an Exabyte 10 H changer before? Running an Exabyte 8500 series tape drive. If so please!!! Tell me how you did it. Christopher A. Los Angeles, Ca.
Re: Getting estimates ..
On Jan 30, 2001, "John R. Jackson" [EMAIL PROTECTED] wrote:
Note that it will not run multiple estimate levels of the same disk at
the same time (I don't think) since that would normally beat on the disk
too hard.
OTOH, if they start all at the same time, they might take advantage of
disk-block buffering within the kernel. It might be worth
experimenting with.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicampoliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist*Please* write to mailing lists, not to me
Re: Calcsize vs gnutar to do estimates
1) Why is gnutar used in stead of calcsize ? For one thing, it knows about exclusion lists. For another, it knows about incrementals. Actually, we might be wrong on both counts. I took a very quick look at the source and there is at least some support for exclusions. And I think it can handle incrementals because it is given multiple timestamps to compare against and appears to add up sizes between the values, which would amount to an incremental. I think. Alexandre Oliva John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
