Hi David,
I don't think I have all the answers to all of your questions, but I did
see a few things that you may want to check into further.
First, I would take a closer look at the chunksize entry in the
Amanda.conf. I think this refers to the size of the files used to hold
your backups. I use a chunk size of 1GB as some filesystems have
difficulty dealing with very large files. This will not reduce the
amount of data per client you can backup. It will simply break the
backup for each entry in the disk list into nice neat chuncks.
Next, please take a look in the /tmp/amanda directory on each troubled
client. In that directory you should have a series of files ending in
".debug". These are log files that will usually contain some hints of
problems. This may help you locate the trouble spot for each client
more efficiently.
If it is at all possible, you may want to consider moving your backup
start time to earlier in the evening. I start my backups at 7:30pm and
they usually finish in the small hours of the morning, depending on how
many level 0's ended up in there.
This may be unrelated to your problem, however I have many solaris
clients on my network I backup as well. I noticed that the TAR and
GZIP/GUNZIP programs that were part of the default install (including
fully patched Solaris 8 systems) were not able to handle files larger
than 2GB. I updated these programs using packages from
www.sunfreeware.com and those problems were fixed.
I know this isn't much information, but hopefully it will help you out
until someone with more experience is able to answer.
Jeff
On Thu, 2002-03-14 at 09:29, David Flood wrote:
> I've asked this question before but I have a litle more information now. My backups
> kick off at 10pm but they are not finishing till 8pm and by that time there is no
> one in the office to change the tape for the next backups starting two hours later.
> This only seems to have taken this long since I upgraded to amanda 2.4.2p2 to avoid
> timeout problems I was having with 2.4.1p1.
>
> Also along with this problem and maybe part of the cause is the fact that
> /export/home/student, staff, grad, msc(the biggest and most important parts of the
> backup - especially staff and student) are not being dumped until afternoon when
> there is a lot of users on the system. Is there a way of forcing amanda to dump and
> tape these during the night when there is no(very little) users on the system? It
> should be noted that /export/blah/blah is contained on the localhost.
>
> I'm using dump and the tapeserver is running Solaris 5.7, I have a tapelist
> containing 20 tapes with a seperate config for monthlys which are run every 4 weeks.
> The tapeserver has 4 network cards qfe0-3, One way of speeding things up was in the
> disklist forcing each host to use the network card for the network it's on, saving
> it from going through the router but I don't know if this is done automatically(my
> networking is not as it should be).
>
> Besides the forcing of amanda to do those dumps during the night is there any tips
> for configuring amanda to do the backups faster? Is there any difference between
> amanda 2.4.2p2 and 2.4.1p1 that would cause this massive backup time difference?
>
> Below I have included a copy of the email sent at the end of the last backup (by the
> way what is that code 3 that ufsdump returned on /export/home/staff). I've also
> attached my amanda.conf which has an include for amanda.conf.dumptypes,
> amanda.conf.net, amanda.conf.tape (Simply to make amanda.conf less cluttered)which
> I've also included below.
>
>
> Email Report:
>
> > These dumps were to tape SCMS19.
> > The next tape Amanda expects to use is: SCMS00.
> >
> > FAILURE AND STRANGE DUMP SUMMARY:
> > vindaloo /export/home lev 0 FAILED [Request to vindaloo timed out.]
> > bicycle /export/home lev 0 FAILED [Request to bicycle timed out.]
> > toast /export/home lev 0 FAILED [Request to toast timed out.]
> > mrwhippy /usr/people lev 0 FAILED [Request to mrwhippy timed out.]
> > localhost /export/home/staff lev 1 FAILED [/usr/sbin/ufsdump returned 3]
> >
> >
> > STATISTICS:
> > Total Full Daily
> > -------- -------- --------
> > Estimate Time (hrs:min) 9:23
> > Run Time (hrs:min) 21:57
> > Dump Time (hrs:min) 16:40 10:36 6:04
> > Output Size (meg) 20635.4 12104.2 8531.2
> > Original Size (meg) 44814.6 25340.4 19474.2
> > Avg Compressed Size (%) 44.4 46.4 41.8 (level:#disks ...)
> > Filesystems Dumped 48 14 34 (1:28 2:3 3:3)
> > Avg Dump Rate (k/s) 352.2 324.9 399.8
> >
> > Tape Time (hrs:min) 1:26 0:50 0:36
> > Tape Size (meg) 20636.9 12104.6 8532.3
> > Tape Used (%) 59.0 34.6 24.4 (level:#disks ...)
> > Filesystems Taped 48 14 34 (1:28 2:3 3:3)
> > Avg Tp Write Rate (k/s) 4104.2 4162.9 4023.6
> >
> >
> > FAILED AND STRANGE DUMP DETAILS:
> >
> > /-- localhost /export/home/staff lev 1 FAILED [/usr/sbin/ufsdump returned 3]
> > sendbackup: start [localhost:/export/home/staff level 1]
> > sendbackup: info BACKUP=/usr/sbin/ufsdump
> > sendbackup: info RECOVER_CMD=/usr/local/bin/gzip -dc |/usr/sbin/ufsrestore -f... -
> > sendbackup: info COMPRESS_SUFFIX=.gz
> > sendbackup: info end
> > | DUMP: Writing 32 Kilobyte records
> > | DUMP: Date of this level 0 dump: Wed Mar 13 17:01:36 2002
> > | DUMP: Date of last level 0 dump: the epoch
> > | DUMP: Dumping /dev/md/rdsk/d5 (arthur:/export) to standard output.
> > | DUMP: Mapping (Pass I) [regular files]
> > ? DUMP: Error in ftw (No such file or directory)
> > | DUMP: The ENTIRE dump is aborted.
> > sendbackup: error [/usr/sbin/ufsdump returned 3]
> > \--------
> >
> >
> > NOTES:
> > planner: Incremental of master:/export/home/mail bumped to level 3.
> > planner: Incremental of roadkill:/export/archive bumped to level 3.
> > planner: Incremental of roadkill:/export/local bumped to level 2.
> > planner: Incremental of localhost:/var bumped to level 2.
> > taper: tape SCMS19 kb 21132160 fm 48 [OK]
> >
> >
> > DUMP SUMMARY:
> > DUMPER STATS TAPER STATS
> > HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s
> > -------------------------- --------------------------------- ------------
> > air -xport/home 1 63 32 50.8 0:02 15.1 0:02 27.0
> > asterix -xport/home 1 5983 1760 29.4 0:17 101.2 0:03 711.6
> > besson -xport/home 1 63 32 50.8 0:02 15.6 0:02 26.9
> > bicycle -xport/home 0 FAILED ---------------------------------------
> > catullus -tullus/ir1 1 127 32 25.2 0:02 18.8 0:03 18.4
> > catullus -tullus/ir2 0 1717151 967136 56.3 43:33 370.1 5:113107.1
> > catullus -tullus/ir3 1 255 32 12.5 0:03 11.4 0:02 27.9
> > catullus -tullus/ir4 0 23008631123392 48.8 62:51 297.9 5:503205.4
> > catullus -tullus/ir5 0 599487 356064 59.4 20:29 289.8 1:164715.6
> > catullus -tullus/ir6 0 1370527 557376 40.7 39:35 234.7 1:574767.9
> > catullus -xport/home 1 31 32 103.2 0:03 11.8 0:02 27.9
> > catullus /opt 0 381375 43328 11.4 4:11 172.9 0:133404.8
> > cluj -xport/home 1 42655 14304 33.5 2:47 85.7 0:043648.6
> > donosti -xport/home 1 63 32 50.8 0:06 5.3 0:02 27.0
> > localhost / 1 255 32 12.5 0:08 3.9 0:03 25.0
> > localhost -dbresearch 1 910815 264480 29.0 13:40 322.5 0:554823.3
> > localhost -/home/grad 0 34369591638624 47.7 50:01 546.0 5:414809.9
> > localhost -t/home/msc 1 67974393055200 44.9 143:56 353.8 10:224911.6
> > localhost -home/staff 1 FAILED ---------------------------------------
> > localhost -me/student 0 126124476292224 49.9 353:51 296.4 21:244900.7
> > localhost -/jumpstart 1 34751992363360 68.0 94:21 417.5 8:044886.8
> > localhost -local/i386 1 895711 321888 35.9 15:23 348.8 1:104629.6
> > localhost -al/oracle7 1 2997983 938400 31.3 26:33 589.1 4:213599.1
> > localhost -ocal/sparc 1 1642719 533440 32.5 22:21 397.8 1:465027.3
> > localhost /opt 3 96831 2624 2.7 0:44 60.2 0:031022.1
> > localhost /usr 1 127 32 25.2 0:29 1.1 0:02 28.1
> > localhost /var 2 126911 11680 9.2 1:03 184.1 0:033578.8
> > locutus -xport/home 1 83999 19360 23.0 2:38 122.6 0:053610.4
> > master / 1 416 416 -- 0:05 78.6 0:04 117.3
> > master -/home/mail 3 687168 687168 -- 20:01 572.0 4:362488.5
> > master /opt 0 109248 109248 -- 5:23 337.8 0:254340.5
> > master /usr 0 464832 464832 -- 20:19 381.2 1:444487.8
> > master /var 0 101696 101696 -- 5:14 323.4 0:293482.7
> > mrwhippy /usr/people 0 FAILED ---------------------------------------
> > obelix -xport/home 1 1503 224 14.9 0:18 12.6 0:03 100.3
> > orac -acle/data1 2 1127935 52768 4.7 7:17 120.8 0:124433.7
> > orac -acle/data2 1 159 32 20.1 0:02 21.3 0:02 26.9
> > orac -acle/data3 1 159 32 20.1 0:06 5.0 0:02 26.8
> > orac -cle/system 0 869183 313856 36.1 13:53 376.8 2:062482.3
> > roadkill / 1 159 32 20.1 0:04 7.5 0:04 17.7
> > roadkill -rt/archive 3 105471 19520 18.5 0:30 653.3 0:101943.6
> > roadkill /export/ftp 1 8735 8064 92.3 0:19 424.9 0:071225.9
> > roadkill -port/local 2 925055 438400 47.4 9:51 742.3 3:252136.1
> > roadkill /export/www 1 4511 2016 44.7 0:24 85.6 0:03 793.1
> > roadkill /opt 0 1088223 126048 11.6 8:24 249.9 0:393194.0
> > roadkill /usr 0 743327 274464 36.9 7:21 621.8 2:221938.5
> > roadkill /var 0 153279 26368 17.2 0:45 581.6 0:211287.4
> > tikka -xport/home 1 63 32 50.8 0:02 20.2 0:02 26.9
> > toast -xport/home 0 FAILED ---------------------------------------
> > turing -xport/home 1 2975 448 15.1 0:34 13.1 0:02 201.2
> > vespasian -xport/home 1 31 32 103.2 0:01 33.1 0:02 26.9
> > vespasian /opt 1 31 32 103.2 0:01 27.9 0:02 27.0
> > vindaloo -xport/home 0 FAILED ---------------------------------------
> >
> > (brought to you by Amanda version 2.4.2p2)
>
> amanda.conf:
>
> ##
> ## amanda.conf - SCMS config file
> ##
>
> org "SCMS" # your organization name for reports
> mailto "support" # space separated list of operators at your site
> dumpuser "root" # the user to run dumps under
>
> inparallel 10 # maximum dumpers that will run in parallel
> netusage 6000 Kbps # maximum net bandwidth for Amanda, in KB per sec
> dumpcycle 4 week # the number of days in the normal dump cycle
> runspercycle 5 # the number of amdump runs in dumpcycle days
> # (4 weeks * 5 amdump runs per week -- just weekdays)
> tapecycle 20 tapes # the number of tapes in rotation
> # 4 weeks (dumpcycle) times 5 tapes per week (just
> # the weekdays) plus a few to handle errors that
> # need amflush and so we do not overwrite the full
> # backups performed at the beginning of the previous
> # cycle
> bumpsize 20 Mb # minimum savings (threshold) to bump level 1 -> 2
> bumpdays 2 # minimum days at each level
> bumpmult 4 # threshold = bumpsize * bumpmult^(level-1)
>
> etimeout 5000 # number of seconds per filesystem for estimates.
> # a positive number will be multiplied by the number of filesystems on
> # each host; a negative number will be taken as an absolute total time-out.
> # The default is 5 minutes per filesystem.
>
> dtimeout 10800 # number of idle seconds before a dump is aborted.
>
> ctimeout 30 # maximum number of seconds that amcheck waits
> # for each client host
>
>
> tapebufs 20
> # A positive integer telling taper how many 32k buffers to allocate.
> # WARNING! If this is set too high, taper will not be able to allocate
> # the memory and will die. The default is 20 (640k).
>
> ##
> ## Tape device
> ##
>
> runtapes 1 # number of tapes to be used in a single run of
> amdump
> #tapedev "/dev/rmt/0bn" # the no-rewind tape device to be used
> tapedev "/dev/rmt/0hbn" # the no-rewind tape device to be used
> # The 'h' specifies a 35GB capacity; should work ok
> # Could also use 'u' and disable software
> compression;
> # Slightly unpredictable (see Amanda FAQ)
> rawtapedev "/dev/null" # the raw device to be used (ftape only)
> tapetype DLT7000 # what kind of tape it is (see tapetypes below)
> labelstr "^SCMS[0-9][0-9]*$" # label constraint regex: all tapes must match
>
> # The DLT isn't a changer; leave these out:
> #tpchanger "chg-manual" # the tape-changer glue script
> #changerfile "/usr/adm/amanda/DailySet1/changer"
> #changerfile "/usr/adm/amanda/DailySet1/changer-status"
> #changerfile "/opt/amanda/etc/amanda/DailySet1/changer.conf"
> #changerdev "/dev/null"
>
> ##
> ## Holding disk
> ##
>
> holdingdisk hd1 {
> comment "main holding disk"
> directory "/export/dumps/amanda" # where the holding disk is
> use -500 Mb # Use all except ~8 GB (leave some for other
>stuff)
> chunksize 8 Gb # size of chunk if you want big dump to be
> # dumped on multiple files on holding disks
> # N Kb/Mb/Gb split disks in chunks of size
> N
> # 0 split disks in INT_MAX/1024 Kb
> chunks
> # -1 same as -INT_MAX/1024 (see
> below)
> # -N Kb/Mb/Gb dont split, dump larger
> # filesystems directly to tape
> # (example: -2 Gb)
> }
>
> ##
> ## Location of log files
> ##
> infofile "/var/adm/amanda/scms/curinfo" # database DIRECTORY
> logdir "/var/adm/amanda/scms" # log directory
> indexdir "/var/adm/amanda/scms/index" # index directory
> tapelist "/var/adm/amanda/scms/tapelist" # list of used tapes
>
> includefile "/opt/amanda/etc/amanda/scms/amanda.conf.tape"
> includefile "/opt/amanda/etc/amanda/scms/amanda.conf.dumptypes"
> includefile "/opt/amanda/etc/amanda/scms/amanda.conf.net"
>
>
> amanda.conf.tape:
>
> ##
> ## Amanda tape config
> ##
> define tapetype DLT {
> comment "DLT tape drives"
> length 20000 mbytes # 20 Gig tapes
> filemark 2000 kbytes # I don't know what this means
> speed 1536 kbytes # 1.5 Mb/s
> }
>
> define tapetype DAT {
> comment "DAT tape drives"
> length 1000 mbytes # these numbers are not accurate
> filemark 100 kbytes # but you get the idea
> speed 100 kbytes
> }
>
> # Definition added by jr
> define tapetype DLT7000 {
> comment "DLT-7000 on E450"
> length 35000 mbytes
> filemark 8 kbytes
> speed 10 mbytes
> }
>
>
> amanda.conf.dumptypes:
>
> # SCMS dump types:
> define dumptype winnt {
> global
> comment "Backup of Windows NT clients using Samba"
> compress server fast
> program "GNUTAR"
> }
> define dumptype scms-static {
> global
> comment "Static filesystems with little change"
> compress client fast
> dumpcycle 10
> priority low
> }
>
> define dumptype scms-static-local {
> global
> comment "Static filesystems with little change on local machine"
> compress client fast
> dumpcycle 10
> priority low
> #holdingdisk no
> }
>
> define dumptype scms-homes {
> global
> comment "Home areas on machines"
> compress client fast
> dumpcycle 5
> priority high
> }
>
> define dumptype scms-homes-local {
> global
> comment "Home areas on local server"
> compress client fast
> priority high
> dumpcycle 5
> #holdingdisk no
> }
>
> define dumptype scms-root {
> global
> comment "Root filesystems"
> compress client fast
> priority medium
> dumpcycle 10
> }
>
> define dumptype scms-root-local {
> global
> comment "Root filesystem on local machine"
> compress client fast
> priority medium
> dumpcycle 10
> #holdingdisk no
> }
>
> define dumptype scms-var {
> global
> comment "var filesystems"
> compress client fast
> priority medium
> exclude "/tmp"
> dumpcycle 10
> }
>
> define dumptype scms-var-local {
> global
> comment "var filesystem on local machine"
> compress client fast
> priority medium
> exclude "/tmp"
> dumpcycle 10
> #holdingdisk no
> }
>
> define dumptype scms-comp {
> global
> comment "Disks to be compressed"
> compress client fast
> }
>
> define dumptype scms-comp-local {
> global
> comment "local disks to be highly compressed"
> compress client fast
> #holdingdisk no
> }
>
> define dumptype scms-nocomp {
> global
> comment "Disks which aren't compressed"
> compress none
> }
>
> define dumptype scms-gnutar {
> global
> program "GNUTAR"
> comment "Disks using GNU tar"
> compress server fast
> }
>
>
> amanda.conf.net:
>
> ##
> ## amanda.conf.net
> ## Network interface definitions
> ##
>
> define interface local {
> comment "a local disk"
> use 10000 kbps
> }
>
> #define interface le0 {
> # comment "10 Mbps ethernet"
> # use 400 kbps
> #}
>
> #define interface hme0 {
> # comment "100 Mbps ethernet"
> # use 4000 kbps
> #}
>
> define interface qfe0 {
> comment "Quad fast ethernet 0"
> use 6000 kbps
> }
>
> define interface qfe1 {
> comment "Quad fast ethernet 1"
> use 6000 kbps
> }
>
> define interface qfe2 {
> comment "Quad fast ethernet 2"
> use 6000 kbps
> }
>
> define interface qfe3 {
> comment "Quad fast ethernet 3"
> use 6000 kbps
> }
>
>
>
--
+----------------------------------------------------------------+
| Jeffery Smith - Systems Administrator - [EMAIL PROTECTED] |
| Sky Computers, Inc. (www.skycomputers.com) |
+----------------------------------------------------------------+
