Taper apparently not working (take 2)
Hello,
My name is Jason. I have been off and on this list for a few years now.
I have just compiled and installed 3.3.2 on my system. I am having trouble
getting taper to work.
When I ran amdump, the system began to correctly pull the data from the
sources, and filling up the holding disk. The output of amstatus always said:
taper status: Searching for a new tape
I can use amtape config label to load tapes from the changer just fine. I
can even use dd to/from the tape drive to verify the correct tape is present.
The output of amcheck config -t shows:
Amanda Tape Server Host Check
-
amcheck-device: Taper::Scan did not quit at
/usr/lib/perl5/site_perl/5.8.8/Amanda/Taper/Scan.pm line 303.
Server check took 0.155 seconds
(brought to you by Amanda 3.3.2)
I have a dell 2750 system with a dell tl4000 tape library, and two LTO4 drives.
For what it's worth, I am backing up a lustre partition mounted on the amanda
server via infiniband. I am using Whamcloud's modified gnu tar to do the
backups. Total storage to backup is 17 terabytes.
Here is my amanda.conf file so you can see my changer definitions:
==begin amanda.conf=
org CLUS-A
mailto my email works
dumpuser backup
inparallel 6
maxdumps 6
dumporder sS
taperalgo largestfit
displayunit g
netusage 54000 Kbps
dumpcycle 0
runspercycle 1
tapecycle 16
etimeout 3600
dtimeout 3600
ctimeout 30
reserve 0
runtapes 15
usetimestamps yes
# maybe change this to all --jason
autoflush yes
#tpchanger chg-zd-mtx
#changerfile /usr/local/etc/amanda/CLUS-A/chg-zd-mtx
#changerdev /dev/sg5
#tapedev /dev/nst1
taper-parallel-write 2
define changer delltl4000 {
tapedev chg-robot:/dev/sg5
property tape-device 0=/dev/nst0 1=/dev/nst1
changerfile /usr/local/etc/amanda/common/chg-robot-dev-sg5
}
tpchanger delltl4000
tapetype LTO4
device_property LEOM TRUE
labelstr ^CLUS-A-[0-9][0-9][0-9][0-9]*$
define tapetype global {
part_size 80G
}
define tapetype LTO4 {
comment Dell LTO4 800Gb - Compression Off
length 802816 mbytes
filemark 0 kbytes
speed 52616 kps
}
holdingdisk hd1 {
comment CLUS-A Backup Server Holding Disk
directory /u1/amanda/CLUS-A
use -1 Gb
chunksize 1 Gb
}
infofile/usr/local/etc/amanda/CLUS-A/curinfo
logdir /usr/local/etc/amanda/CLUS-A/log
indexdir/usr/local/etc/amanda/CLUS-A/index
tapelist/usr/local/etc/amanda/CLUS-A/tapelist
#_ or - are acceptable
# but this value is no longer ok. Question: is the default oK?
#amrecover_changer chg-zd-mtx
define dumptype global {
comment Global Definitions
program GNUTAR
auth bsdtcp
compress none
dumpcycle 0
index yes
record yes
exclude list /usr/local/etc/amanda/CLUS-A/exclude
maxdumps 6
estimate server
}
define dumptype full-sb-gztar {
global
comment Full Backups Server Best Compression
compress server best
priority high
}
define dumptype full-cb-gztar {
global
comment Full Backups Client Best Compression
compress client best
priority high
}
define application-tool app_lustretar {
plugin amgtar
property ATIME-PRESERVE NO
property GNUTAR-PATH /usr/local/bin/tar
property SPARSE YES
property ACLS NO
property SELINUX NO
property XATTRS YES
property EXIT-HANDLING 1=GOOD 2=BAD
# change a default NORMAL regex to a STRANGE regex.
property STRANGE : socket ignored$
# add three new IGNORE regex
property IGNORE : Directory is new$
property append IGNORE : Directory has been renamed
property append IGNORE file changed as we read it$
}
define application-tool app_amstar {
comment amstar
plugin amstar
}
define dumptype dt_sbc_lustretar {
comment lustre modified Tar
global
program APPLICATION
application app_lustretar
compress server best
#dumpcycle 0
index yes
record yes
maxdumps 6
estimate server
}
define dumptype dt_sbc_amstar {
comment STAR Server Based Compression Definitions
program APPLICATION
application app_amstar
auth bsdtcp
compress server best
dumpcycle 0
index yes
record yes
maxdumps 6
estimate server
}
define dumptype dt_cbc_amstar {
comment STAR Client Based Compression Definitions
program APPLICATION
application app_amstar
auth bsdtcp
compress client best
dumpcycle 0
index yes
record yes
maxdumps 6
estimate server
}
# for ethernet I used 80% of the theoretical max bw
# thus for gigabit ethernet at 125 megabytes per second, its 100 megabytes
# or 10 kilobytes per second
# for infiniband:
#40bps ~= 4Gbytes per second
#4000 megabytes per second
#400 kilobytes per second
#* .8 == 320 kilobytes per second at 80%
define interface local {
comment Local Disk
# stupid number, but who cares...
use 5400 Kbps
}
define interface lo {
comment Loopback Device
# stupid number, but who cares...
Re: amrecover and tape seeking
Good question. I would like to know as well. As for me, I pretty much always pre-position to the file number with mt. I hadn't realized amanda didn't do that. Chris Miller wrote: amadmin find shows a record of the location on tape (record number) where each dump record exists. Given this information, why does amrecover start at the beginning of the tape during a restore and read each record sequentially on the tape until it finds the correct record? I can significantly reduce restore times by suspending amrecover and using mt fsf n to pre-position the tape one or two records before the record it wants. I gather the code is doing a pattern match for hostname, diskname, and date. Why not use this information to get the record number and position the tape appropriately? Currently I average 150 records per DLT 40/80 and it takes forever to do a restore, particularly when multiple records are needed from the same tape and amrecover rewinds the tape after each restore unnecessarily given the newer dumps are at the end of the tape. (yes, I am using the no-rewind device :-). I realize it's safer to start from the beginning of the tape, but after a few all nighters down at the data center over the years, this is a question that's plagued me given the several versions released. Just want to understand the logic or respectfully request an enhancement. Kind regards, Chris -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 Email to: [EMAIL PROTECTED] Twiki: http://twiki.wrs.com/do/view/Main/JasonBrooks Senior Systems Administration Analyst Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: amrecover and tape seeking
Hello, I found this in the amanda.conf.in file. It's new to me. I have been using my same amanda.conf file since 2.4.2p1, so go figure. amrecover_do_fsf yes# amrecover will call amrestore with the # -f flag for faster positioning of the tape. amrecover_check_label yes# amrecover will call amrestore with the # -l flag to check the label. amrecover_changer @DEFAULT_TAPE_DEVICE@# amrecover will use the changer if you restore # from this device. # It could be a string like 'changer' and # amrecover will use your changer if you # set your tape with 'settape changer' Jason Brooks wrote: Good question. I would like to know as well. As for me, I pretty much always pre-position to the file number with mt. I hadn't realized amanda didn't do that. Chris Miller wrote: amadmin find shows a record of the location on tape (record number) where each dump record exists. Given this information, why does amrecover start at the beginning of the tape during a restore and read each record sequentially on the tape until it finds the correct record? I can significantly reduce restore times by suspending amrecover and using mt fsf n to pre-position the tape one or two records before the record it wants. I gather the code is doing a pattern match for hostname, diskname, and date. Why not use this information to get the record number and position the tape appropriately? Currently I average 150 records per DLT 40/80 and it takes forever to do a restore, particularly when multiple records are needed from the same tape and amrecover rewinds the tape after each restore unnecessarily given the newer dumps are at the end of the tape. (yes, I am using the no-rewind device :-). I realize it's safer to start from the beginning of the tape, but after a few all nighters down at the data center over the years, this is a question that's plagued me given the several versions released. Just want to understand the logic or respectfully request an enhancement. Kind regards, Chris -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 Email to: [EMAIL PROTECTED] Twiki: http://twiki.wrs.com/do/view/Main/JasonBrooks Senior Systems Administration Analyst Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
RE: Amanda not backing up Veritas file systems correctly
You know, I am always amazed at the different things to be found in amanda's source code. I tried writing a more detailed INTERNALS document a while ago: it did not work out. I will try the make getfsent. thank you again for your time. --jason ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 Email to: [EMAIL PROTECTED] Twiki: http://twiki.wrs.com/do/view/Main/JasonBrooks Senior Systems Administration Analyst Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008 -Original Message- From: Paul Bijnens [mailto:[EMAIL PROTECTED] Sent: Friday, 28 May, 2004 00:02 To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: Amanda not backing up Veritas file systems correctly [EMAIL PROTECTED] wrote: Actually, I don't have those entries in /etc/vfstab. I will give that a try. For future reference, in order to handle transient filesystems such as this, it might be better off evaluating the output of mount rather than the filesystem table. Except I can't remember if it gives FS info. If you have amanda sources around, take a look in the dir client-src. There is file name getfsent.c which contains the code. The beginning comment is: /* * You are in a twisty maze of passages, all alike. * Geesh. */ And that becomes clear when trying to understand the next few lines :-) To help debugging this, you can compile getfsent into a standalone program. Just type make getfsent in that directory. When you run it as ./getfsent it shows which blocks were defined in the maze of passages and shows what amanda thinks about each filesystem on the host. -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] ** * * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, F6, * * quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ...* * ... Are you sure? ... YES ... Phew ... I'm out * ** *
Amanda not backing up Veritas file systems correctly
Hello, I have been using amanda for nearly two years now. Most things seem pretty easy to figure out... At least if you like source code and cscope... :) I currently am running amanda-2.4.4p2 on both server and client, the server being a redhat 7.x system. The client is a recently added solaris 9 server running clearcase. Its /, /bybee1, and /var filesystems are using ufs. The filesystems /bybee[2..10] are using vxfs. Amanda can back them up ok, but if I mount their snapshots separately then try to backup the snapshots, they are mistaken for ufs filesystems and ufsdump therefore dies. Ok, more information: The filesystems I am testing against are: Filesystemkbytesused avail capacity Mounted on /dev/dsk/c1t0d13s2 51423232 29078 48182027 1%/bybee9 /dev/dsk/c1t0d18s2 51423232 29078 48182020 1%/bybee9SnapVol /bybee9 is simply a vxfs filesystem mounted at /bybee9 /bybee9SnapVol is a snapshot of /bybee9 mounted at /bybee9SnapVol I am running planner directly on the server: planner Daily2 bybee /bybee9SnapVol planner Daily2 bybee /dev/dsk/c1t0d18s2 If I run planner Daily2 bybee /bybee9, I get a successful run of vxdump (looking in /tmp/amanda/sendsize) However, if I run the planner on either /bybee9SnapVol, or its device file, sendsize tries to run ufsdump instead. If I use the solaris program fstyp, it returns that the device files are vxfs filesystems. Where is amanda going wrong (or me for that matter)? *** * disklist entries I created for my tests * *** bybee /bybee9SnapVolnocomp-root 3 bybee /bybee9 nocomp-root 3 bybee /dev/dsk/c1t0d18s2 nocomp-root 3 * failed sendbackup debug file * sendsize: debug 1 pid 8725 ruid 11600 euid 11600: start at Thu May 27 12:18:00 2004 sendsize: version 2.4.4p2 sendsize[8728]: time 0.004: calculating for amname '/bybee9SnapVol', dirname '/bybee9SnapVol', spindle 3 sendsize[8728]: time 0.005: getting size via dump for /bybee9SnapVol level 0 sendsize[8725]: time 0.005: waiting for any estimate child: 1 running sendsize[8728]: time 0.009: calculating for device '/bybee9SnapVol' with '' sendsize[8728]: time 0.009: running /usr/sbin/ufsdump 0Ssf 1048576 - /bybee9SnapVol sendsize[8728]: time 0.010: running /usr/local/libexec/killpgrp-2.4.4p2 sendsize[8728]: time 0.021: Unable to create temporary directory in any of the directories listed below: sendsize[8728]: time 0.022: /tmp/ sendsize[8728]: time 0.023: /var/tmp/ sendsize[8728]: time 0.023: / sendsize[8728]: time 0.024: Please correct this problem and rerun the program. sendsize[8728]: time 0.024: DUMP: `/bybee9SnapVol' is not on a locally mounted filesystem sendsize[8728]: time 0.025: DUMP: The ENTIRE dump is aborted. sendsize[8728]: time 0.025: . sendsize[8728]: estimate time for /bybee9SnapVol level 0: 0.016 sendsize[8728]: no size line match in /usr/sbin/ufsdump output for /bybee9SnapVol sendsize[8728]: . sendsize[8728]: estimate size for /bybee9SnapVol level 0: -1 KB sendsize[8728]: time 0.026: asking killpgrp to terminate sendsize[8728]: time 1.019: done with amname '/bybee9SnapVol', dirname '/bybee9SnapVol', spindle 3 sendsize[8725]: time 1.020: child 8728 terminated normally sendsize: time 1.020: pid 8725 finish time Thu May 27 12:18:01 2004 ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 Email to: [EMAIL PROTECTED] Twiki: http://twiki.wrs.com/do/view/Main/JasonBrooks Senior Systems Administration Analyst Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
RE: Amanda not backing up Veritas file systems correctly
Actually, I don't have those entries in /etc/vfstab. I will give that a try. For future reference, in order to handle transient filesystems such as this, it might be better off evaluating the output of mount rather than the filesystem table. Except I can't remember if it gives FS info. Thanks for the reply... --jason ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 Email to: [EMAIL PROTECTED] Twiki: http://twiki.wrs.com/do/view/Main/JasonBrooks Senior Systems Administration Analyst Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008 -Original Message- From: Paul Bijnens [mailto:[EMAIL PROTECTED] Sent: Thursday, 27 May, 2004 14:50 To: [EMAIL PROTECTED] Cc: [EMAIL PROTECTED] Subject: Re: Amanda not backing up Veritas file systems correctly [EMAIL PROTECTED] wrote: Its /, /bybee1, and /var filesystems are using ufs. The filesystems /bybee[2..10] are using vxfs. Amanda can back them up ok, but if I mount their snapshots separately then try to backup the snapshots, Amanda gets if fstype from the /etc/vfstab file, or defaults to the DUMP (from configure --with-dump=XXX). Is /bybee9SnapVol in that file? (maybe with noauto for mounting?) -- Paul Bijnens, XplanationTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] ** * * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, F6, * * quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ...* * ... Are you sure? ... YES ... Phew ... I'm out * ** *
Negative numbers seen in amreport stats
Hello, I am seeing negative numbers in my amanda stats after a dump. I have faith the backup is occuring. It looks like when the backup size is small enough, or occurs quickly enough, the numbers calculated for dumper-times, or taper-times (and their associated rates) turn out negative. I have compared the values in my amreport to those in the amdump and log files: they match (with some rounding errors). I can think of two reasons this might be: 1) cumulative rounding errors 2) the result of write caching on the part of the disk and/or tape drivers and electronics. I am not alarmed by this, but wanted to tell you all about it in case it's a bug. I am using a dual processor linux system as the amanda server with a quantum DLT8000 tape drive (part of a quantum/atl tape changer), and two ide drives (7200 rpm) using linux software raid in linear mode for my holding space. This server is running amanda-2.4.4p2. I will include some of the stats lines here: For what it's worth, all of the machines listed below are redhat9 except for the following: weed, crater: solaris8 fanno: solaris7 methane: solaris2.5.1 All client machines are running amanda-2.4.4p2 DUMP SUMMARY: DUMPER STATSTAPER STATS HOSTNAME DISK L ORIG-KBOUT-KB COMP% MMM:SS KB/s MMM:SS KB/s --- --- - bayou/1 310 310 --0:06 50.4 0:01 338.6 bayou/bayou1 1 390 390 --0:13 29.6 0:-2 -366.6 bespin /1 870 870 --0:02 555.9 0:-3 -382.8 brine/boot12020 --0:00 186.6 0:-3 -5.5 brine/brine1 1 390 390 --0:15 26.1 0:-2 -145.5 cattail /1 1140 1140 --0:00 85869.2 0:-3 -322.8 cattail /var 1 8150 8150 --0:11 739.6 0:-2 -5498.4 crater /crater1 1 128 128 --0:09 14.2 0:-3 -55.3 cricket /cricket11 390 390 --0:429.2 0:-1 -237.5 cricket /var 1 4630 4630 --0:04 1098.2 0:-2 -1618.8 critter hda1 11010 --0:00 40.0 0:-4 -2.9 critter hda5 1 11290 11290 --0:20 554.3 0:-2 -7822.3 fanno/fanno1 1 21850 21850 --0:32 677.8 0:-1 -300339.5 fanno/fanno1/releases 1 10540 10540 --2:59 59.0 0:-1 -6833.1 lily /1 400 400 --0:-2 -336.4 0:-1 -29099.4 lily /boot12020 --0:01 28.0 0:-2 -16.8 lily /lily1 1 240 240 --0:327.4 0:-4 -70.3 methane c0t0d0s3 13232 --0:00 -49.5 0:08 3.9 methane c0t0d0s6 13232 --0:074.3 0:-2 -23.6 weed /weed1/images1 470 470 --0:01 854.6 0:-2 -462.8 ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 Email to: [EMAIL PROTECTED] Twiki: http://twiki.wrs.com/do/view/Main/JasonBrooks Senior Systems Administration Analyst Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
RE: nfs mounted executable
Hello,
I assume you mean the client side programs? That is to say, inetd calls amandad on an nfs mount?
If so, setting of an environment variable for your library search path might do it. If this works,
the only mod you will need to do to your system is to change your xinetd.conf file. But you had to
do that anyway to make amanda work... You really want to use shared libraries.
Please see ld.so(8) -- man page
- run ldd on the executable to find out what libraries it's looking for.
- set LD_LIBRARY_PATH=insert path name here to libraries
- export this variable to the environment so it's not just a shell variable
- run ldd again, to see if the library is now found. It should be.
According to redhat 9's man page on ld.so, several things are searched for the libaries in question.
First, the executable itself might know precisely where the libraries are located. second the
paths in LD_LIBRARY_PATH are searched, then third the /etc/ld.so.cache file is searched. You don't
want to add this to the cache file as EVERY library search done by the system for every executable
will search your nfs mounts.
it is not clear if xinetd will source any of the shell configs, therefore it's not likely you can
just put this environment variable in your backup user's .bashrc or .cshrc. with xinetd you can use
the following sample entry. Please note, this is from the xinetd man page: I haven't tested it.
Please also note though: the man page states you can't have any spaces in your variable.
service amanda
{
socket_type = dgram
protocol= udp
wait= yes
user= bv-amanda
group = bv-backup
server =
/net/jetstream/jetstream1/amanda-2.4.4p2client/libexec/amandad
env =
LD_LIBRARY_PATH=/net/jetstream/jetstream1/amanda-2.4.4p2client/lib
}
~~~
Jason Brooks ~ (503) 641-3440 x1861
Direct ~ (503) 924-1861
Email to: [EMAIL PROTECTED]
Twiki: http://twiki.wrs.com/do/view/Main/JasonBrooks
Senior Systems Administration Analyst
Wind River Systems
8905 SW Nimbus ~ Suite 255
Beaverton, Or 97008
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of DK Smith
Sent: Wednesday, February 18, 2004 2:30 PM
To: [EMAIL PROTECTED]
Subject: nfs mounted executable
Hello all,
I was experimenting with a different installation strategy...
That is, in order to not install amanda on every host... I am
attempting to place a single amanda installation on a file
server; making it available via an NFS share. The nfs-clients
of the same architecture (in this case Linux RH8.0) mount
this NFS share and run the amanda programs... (the tape
server and the amanda clients).
The first problem I hit when running amcheck on the tape
server was that of the shared libraries. (so this is quickly
turning into NOT an amanda question grin)
amcheck: error while loading shared libraries:
libamserver-2.4.4p2.so:
cannot open shared object file: No such file or directory
Maybe someone that has been down this road before can tell
me, stop now as this is not practical, or something
similar? Or is there a config option to not build shared
libraries and hence another host would not look for them at
the same hard-coded path (in which the particular binary was built)
thanks!
DK
PS: I am not intending to back-up NFS mounted data.
RE: Negative numbers seen in amreport stats
Ah, it's funny you should ask: Yesterday afternoon I just synched all of the clocks. I do have a problem with clock drift, even with xntp. I think one of the upstream servers is skewed, but i have no choice about that. I have toyed with the idea of getting my own radio clock and becoming my own stratum 1 server. However, that's beyond this maiing list. The irony here is that the most recent backup is where I got the negative numbers was started 3 hours after I had synced the clocks. I will keep an eye on the clocks but they may not be the issue. Thank you for your swift reply... --jason ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 Email to: [EMAIL PROTECTED] Twiki: http://twiki.wrs.com/do/view/Main/JasonBrooks Senior Systems Administration Analyst Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008 -Original Message- From: Gerhard den Hollander [mailto:[EMAIL PROTECTED] Sent: Thursday, February 19, 2004 11:51 AM To: Jason Brooks Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED] Subject: Re: Negative numbers seen in amreport stats * Jason Brooks [EMAIL PROTECTED] (Thu, Feb 19, 2004 at 10:40:31AM -0800) Hello, I am seeing negative numbers in my amanda stats after a dump. I have This may be barking up the wrong tree, but are you sure all your systems have a correct clock ? And one that is being synchronised (xntpd is your friend) to the same refrence clock ? Even if it is not causing this problem, if the clocks ont eh machines in your network are out of sync, you will eventually encounter really wierd problems. And if you use rdate/ntpdate to sync (which bruteforces the clock to the corect time, even if this means moving it minutes or hours at once) you will definitely get weird errors and very hard to trace porblems. (XFree w/ motif doesn;t like it at all, not all NFS implementations like it if a files is finished writing ``before'' it is actually created. Commercial software with a strict license manager hates it ) So: login to variosu machines and run the command date. If the numbers are off by more then a second, please consider using xntpd on your machines. Kind regards, -- Gerhard den Hollander Phone :+31-10.280.1515 ICT manager Direct:+31-10.280.1539 Fugro-Jason Fax :+31-10.280.1511 [EMAIL PROTECTED] POBox 1573 visit us at http://www.Fugro-Jason.com 3000 BN Rotterdam JASON...#1 in Reservoir CharacterizationThe Netherlands This e-mail and any files transmitted with it are confidential and intended solely for the use of the addressee. This e-mail shall not be deemed binding unless confirmed in writing. If you have received it by mistake, please let us know by e-mail reply and delete it from your system; you may not copy this message or disclose its contents to anyone. Please note that any views or opinions presented in this e-mail are solely those of the author and do not necessarily represent those of the company. E-mail transmission cannot be guaranteed to be secure or error-free. The sender therefore does not accept liability for any errors or omissions in the contents of this message, which arise as a result of e-mail transmission.
Re: The next tape Amanda expects to use is: a new tape
Hello, I am going to guess on this one. My first thought is that there is now a gap in your tapelist. amanda wanted tape number 4. there is nothing there now, which might be why amanda sez a new tape in your messages. I believe the order of your tapelist dictates to amanda what the next tape should be. the last entry is the next tape. Assuming you have entered a new tape number 4, you might ensure that is the last entry, and the first field (date field) is the oldest date (the number zero ought to do it!) you can use amcheck to test. It's unfortunate that your tape drive apparently has no end of tape detection mechanism. on a number of them it's a series of holes passing an optical detector. If your tape drive has these, then they might be full of dust, or something. Since the tape broke, I am going to assume you cannot repair it, and that everything successfully flushed to tape was lost. This info ought to be in your logs. you might need to force these backups to be redone. I hope this works out well for you... --jason John wrote: Long story short. Ran tapetype on a 230m tape. Then got a box of 170m tapes without noticing. Tape 4, a 170M tape was snapped at it's end. I amrmtape tape4, and now amanda insists it needs a new tape to flush the stored backups to. I'm now gettings: *** A TAPE ERROR OCCURRED: [cannot overwrite active tape Daily01]. Some dumps may have been left in the holding disk. Run amflush to flush them to tape. The next tape Amanda expects to use is: a new tape. Is there a way to make it use the existing 7 remaining tapes instead of throwing a new tape at it? It's mostly a time thing for getting a new tape, but it'd be nice to be able to go on with the next tape in the cycle. Any suggestions? I google'd for this and the only useful thing relating to this is that 'it's a showstopper' which is not what I really want to hear. :( If it IS a showstopper, I'll have to hope the dump to disk has enough space til the tape can arrive. Thanks folks. j -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 Email to: [EMAIL PROTECTED] System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: Help with AMRESTORE
Hello, if you are using amrestore to grab the image from tape, then: cd working directory amrestore -p [and other options] | tar -xpf - the -p option in amrestore outputs the datastream to a pipe (stdout). this sets up the buffering correctly. the -p option in tar will restore the files with the original dates and permissions. if you know already the location of the image on tape, then A) position tape mt -f non-rewinding tape dev fsf filenumber B) run dd if=/dev/st1 count=1 bs=1024b of=test C) read file test. note the first three lines will say something like: AMANDA: FILE 20021219 jetstream hda1 lev 1 comp N program /sbin/dump To restore, position tape at start of file and run: dd if=tape bs=32k skip=1 | /sbin/restore -f... - this last line might read something like: dd if=tape bs=32k skip=1 | tar -xpf - ... --jason On Fri, Dec 20, 2002 at 11:26:07AM -0500, wab wrote: Greetings. I'm attempting to recover a filesystem which is about 15-18 gigs in size. Amanda is using Tar to compress its archives... I tried to restore the image to a filesystem large enough to handle the data, before moving back to the original location; but I got a File size limit exceeded. (It reached about 2.5 gigs in size). I understand it's possible to Pipe the restore through TAR to de-compress it back to regular files; but I can't figure out the syntax exactly; has anybody ever done this? I think that's what I must do, either that or find a filesystem/operating system that can handle giant huge files. I'll provide version numbers of the OS/amanda if necessary... Thanks Wayne Byarlay ITD, Purdue Libraries [EMAIL PROTECTED] -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: cant switch to incremental dump
Hello, Am I reading you correctly that you haven't gotten a full dump yet and amanda is displaying 'can't switch to incremental dump'? If this is the case, read on... Do you have the parameter reserve set in your amanda.conf file? if not, it defaults to 100%. this means that the holding disk is 100% reserved for incremental backups. you get the aforementioned message because amanda hasn't performed a full backup yet, therefore there isn't a such thing as an incremental. I have my reserve value set to 30. My thinking (possibly flawed) is if I leave enough non-reserved holding space to accomidate a number of full backups, I will be ok. At a minimum, I must ensure that 70% of my holding space can hold the largest level 0 backup I have to do. On Fri, Dec 20, 2002 at 09:16:54AM -0500, [EMAIL PROTECTED] wrote: Greetings, (I sent a mail about this a few days ago, so if I am duplicating info please forgive) I still cannot get one of my disks to dump correctly. It is reporting the error can't switch to incremental dump. I have unforced the dump w/ amadmin just in case. I have a tape with more then enough space to archive this data. I have plenty of space in my holding disk. I just cannnot figure out why this dump is trying to do an incremental dump and why it cannot. If I use amadmin and force a level 0, shouldn't amanda only run a level 0? The logs in /tmp/amanda dont point me to the solution either. I am really confused here and any help would be appreicated. Thank You. Andrew -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Peer review for chg-scsi usage document
Hello, I am in the process of documenting in a bit more detail chg-scsi. I would love it if you all would tell me what you thought so far. I would eventually like to expand it to include some of amanda's actual interactions, etc. --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008 chg-scsi.doc Description: MS-Word document
Re: Peer review for chg-scsi usage document
Hello, You know, I think I typed the .doc extension by habit: It's a plain text file. I used vim to type it. Sorry for the mixup. Were there actually any headers in it other than the name of the file? On Thu, Jul 25, 2002 at 01:05:02PM -0700, Anthony A. D. Talltree wrote: I would love it if you all would tell me what you thought so far Encrypting the document makes it kinda tough to read. -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: chg-chio and barcode's
Hello, I have just had this problem myself with chg-scsi not seeing my barcodes. I am working on a fix for this. To determine if your changer and barcode reader are functioning correctly, grab and compile mtx. You also can use mtx with chg-zd-mtx. --jason On Fri, Jun 28, 2002 at 03:16:31PM +0200, [EMAIL PROTECTED] wrote: Has anyone written barcode label support into chg-chio ? The library I've got here (HP Surestore 2x28 DLT) does not want to play with chg-scsi (on FreeBSD 4.5). The latter throws a coredump on the sense error messages. chg-chio works perfectly fine - but rather than use the barcode/labels - it actually needs to look on the tape for the labels. Which seems a bit of a pity. Thanks ! Dw. -- Dirk-Willem van Gulik -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: SMBCLIENT program not available
hello, Did you ever get this resolved? I just remembered also that if samba isn't setup correctly, this might produce this error. I believe in order for samba to work at all, it needs an smb.conf file in the appropriate location. --jason On Tue, Jun 25, 2002 at 11:25:26PM +0200, Martin Schmidt wrote: Hi, I am just testing amanda on my privat network before I will install it in the company. I have a SuSE-Linux 7.3 (rappelkasten) with samba, amanda and tape, and for testing a Win-Nt4 (rumpl). There is no problem running amanda on the linux. So I tried to backup a share from the NT. I created a user amanda on the NT with all privileges on one share. When I do a smbclient from commandline I can see the contents of that share. When I start amcheck, I get the messages: Amanda Backup Client Hosts Check ERROR: localhost: [This client is not configured for samba: //rumpl/lwd] ERROR: localhost: [SMBCLIENT program not available] Client check: 1 host checked in 0.084 seconds, 2 problems found (brought to you by Amanda 2.4.2p2) -- First i supposed, amanda is not configured with-smbclient when I install it from SuSE's CDs, so I got the file from source-forge and did a configure with --with-smbclient=/usr/bin/smbclient --with-user=amanda --with-group=disk (smbclient _is_ there, user and group I took from SuSE's default ). The messages are the same. There is no hint in the FAQs, the doc - directory, but several questions in the mailing list, but all unanswered. So I hope, I get an answer and wipe this problem off. Thanks Martin -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Compile error, resolution, confusion
Hello, I have been reading through the archives and found a bit of information that helped me. for amanda 2.4.3b3, I received errors while compiling on a solaris8 (bybee) machine. I applied a change I found in reference to this error, and the compiler error went away. the wierd part: this error didn't happen on different machines running solaris8(bramble), 7, 251, redhat 6.2-7.2, or freebsd3.4. the change was to line 244 of client-src/sendbackup-dump.c: I changed no_record to options-no_record. Interestingly enough, this appears to compile just as well on the other machines too. I have installed on both solaris machines: binutils-2.11.2-sol8-sparc-local bison-1.28-sol8-sparc-local flex-2.5.4a-sol8-sparc-local gcc-3.0.3-sol8-sparc-local libgcc-3.0.3-sol8-sparc-local make-3.79.1-sol8-sparc-local I will include the compiler error messages below. Please note that the biggest difference between the compiling and non-compiling machines is that the non-compiling machine has a slightly older os patch level, Veritas File system and Veritas Volume manager installed, and is a clearcase server. Otherwise, the gcc, binutils, make, etc packages are all the same. What do you all think? --jason the error is: gcc -DHAVE_CONFIG_H -I. -I. -I../config -I../common-src -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -g -O2 -c `test -f sendbackup-dump.c || echo './'`sendbackup-dump.c sendbackup-dump.c: In function `start_backup': sendbackup-dump.c:244: `no_record' undeclared (first use in this function) sendbackup-dump.c:244: (Each undeclared identifier is reported only once sendbackup-dump.c:244: for each function it appears in.) make[1]: *** [sendbackup-dump.o] Error 1 make[1]: Leaving directory `/net/jetstream/jetstream1/amanda/bybee/client-src' make: *** [all-recursive] Error 1 -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: use of changer
Hello, Here is the output from mtx and loaderinfo. Note that though loaderinfo doesn't show a barcode reader, mtx's output clearly shows a barcode reader. I can confirm that those VolumeTags are accurate. --mtx--- [streak : changer]$ mtx inquiry Product Type: Medium Changer Vendor ID: 'ATL ' Product ID: 'L500 632' Revision: '0026' Attached Changer: No [streak : changer]$ mtx status Storage Changer /dev/changer:1 Drives, 13 Slots ( 0 Import/Export ) Data Transfer Element 0:Full (Storage Element 1 Loaded) Storage Element 1:Empty Storage Element 2:Full :VolumeTag=RBN066 Storage Element 3:Full :VolumeTag=RBN067 Storage Element 4:Full :VolumeTag=RBN068 Storage Element 5:Full :VolumeTag=RBN069 Storage Element 6:Full :VolumeTag=RBN070 Storage Element 7:Full :VolumeTag=RBN071 Storage Element 8:Full :VolumeTag=RBN072 Storage Element 9:Full :VolumeTag=RBN073 Storage Element 10:Full :VolumeTag=RBN074 Storage Element 11:Full :VolumeTag=RBN075 Storage Element 12:Full :VolumeTag=RBN076 Storage Element 13:Full :VolumeTag=RBN077 -loaderinfo [streak : jason0]$ loaderinfo -f /dev/sg1 Product Type: Medium Changer Vendor ID: 'ATL ' Product ID: 'L500 632' Revision: '0026' Attached Changer: No Bar Code Reader: No EAAP: Yes Number of Medium Transport Elements: 1 Number of Storage Elements: 13 Number of Import/Export Element Elements: 0 Number of Data Transfer Elements: 1 Transport Geometry Descriptor Page: Yes Invertable: No Device Configuration Page: Yes Can Transfer: No On Fri, Jun 21, 2002 at 01:25:05PM -0400, Nicholas Berry wrote: Hi. The sense data says a SCSI device/bus or power on reset occurred. All is not well. I assume you're using Linux, from the /dev/sg1 reference. Get the mtx package installed, and see what that makes of your changer. Nik Jason Brooks [EMAIL PROTECTED] 06/20/02 06:36PM Hello, How do I get amanda to use the barcode reader on my brandnew scsi-changer? (it's an independant changer with its own scsi-generic interface) the changer appears to work otherwise, though I haven't actually begun testing backups on it... the ./docs/TAPE.CHANGER file contains the following: To initialize the label mapping you can also do an amtape config show. This will put the mapping for all tapes in the magazine into the mapping file. This doesn't work for me. enclosed is my changer.conf file and the relevent entries from my amanda.conf file. I would like to have amanda cross-reference barcodes to tape labels, perhaps with an inventory control to verify barcode/amlabel mapping. if amand needs a certain label, it tells the chager to load barcode #n. //cut from amanda.conf tpchanger chg-scsi tapedev 0 # the abstract changer config in the changer file changerfile /usr/local/etc/amanda/changer/changer.conf //my changer.conf contents: number_configs 1 eject 0 sleep 140 cleanmax 20 changerdev /dev/sg1 config 0 drivenum0 dev /dev/nst1 startuse0 havebarcode 1 enduse 12 cleancart 13 statfile/var/amanda/DailySet2/tape0-slot cleanfile /var/amanda/DailySet2/tape0-clean usagecount /var/amanda/DailySet2/totaltime tapestatus /var/amanda/DailySet2/tapestatus labelfile /usr/local/etc/amanda/changer/labelfile //sample contents of a chg-scsi*debug file chg-scsi: debug 1 pid 1218 ruid 11600 euid 11600 start time Wed Jun 19 15:27:48 2002 chg-scsi: $Id: chg-scsi.c,v 1.1 2002/06/18 18:25:41 jason0 Exp $ ARG [0] : /usr/local/libexec/chg-scsi ARG [1] : -slot ARG [2] : current Number of configurations: 1 Tapes need eject: No Inv. auto update: No barcode reader : Yes Tapes need sleep: 140 seconds Cleancycles : 20 Changerdevice : /dev/sg1 Labelfile : /usr/local/etc/amanda/changer/labelfile Tapeconfig Nr: 0 Drivenumber : 0 Startslot : 0 Endslot : 12 Cleanslot : 13 Devicename: /dev/nst1 changerident : none SCSITapedev : none tapeident : none statfile : /var/amanda/DailySet2/tapestatus Slotfile : /var/amanda/DailySet2/tape0-slot Cleanfile : /var/amanda/DailySet2/tape0-clean Usagecount: /var/amanda/DailySet2/totaltime # START SenseHandler Ident = [L500 632], function = [generic] # START GenericSenseHandler # START DecodeSense GenericSenseHandler : Sense Keys Extended Sense ASC 29 ASCQ 00
attempt to use chg-zd-mtx...
Hello again, I created another amanda configuration and set it up to use chg-zd-mtx since mtx apparently showed accurate information. Here is the result of just running chg-zd-mtx -info -- none could not determine current slot and here is the chg-zd-mtx*debug file chg-zd-mtx: debug 1 pid 16486 ruid 11600 euid 11600 start time Fri Jun 21 15:42: 55 2002 15:42:55 Arg info: $# = 1 $0 = /usr/local/libexec/chg-zd-mtx $1 = -info 15:42:55 Running: /usr/local/sbin/mtx status 15:42:55 Exit code: 0 Stdout: Storage Changer /dev/sg1:1 Drives, 13 Slots ( 0 Import/Export ) Data Transfer Element 0:Full (Storage Element 1 Loaded) Storage Element 1:Empty Storage Element 2:Full :VolumeTag=RBN066 Storage Element 3:Full :VolumeTag=RBN067 Storage Element 4:Full :VolumeTag=RBN068 Storage Element 5:Full :VolumeTag=RBN069 Storage Element 6:Full :VolumeTag=RBN070 Storage Element 7:Full :VolumeTag=RBN071 Storage Element 8:Full :VolumeTag=RBN072 Storage Element 9:Full :VolumeTag=RBN073 Storage Element 10:Full :VolumeTag=RBN074 Storage Element 11:Full :VolumeTag=RBN075 Storage Element 12:Full :VolumeTag=RBN076 Storage Element 13:Full :VolumeTag=RBN077 15:42:55 SLOTLIST - firstslot set to 2 15:42:55 SLOTLIST - lastslot set to 13 15:42:55 Config info: firstslot = 2 lastslot = 13 cleanslot = -1 cleancycle = -1 offline_before_unload = 0 unloadpause = 0 autoclean = 0 autocleancount = 99 havereader = 0 driveslot = 1 poll_drive_ready = 3 max_drive_wait = 120 15:42:55 SETUP- current slot -1 less than 2 ... resetting to 2 15:42:55 Exit (2) - none could not determine current slot chg-zd-mtx: pid 16615 finish time Fri Jun 21 15:42:56 2002 What do you think of this? On Fri, Jun 21, 2002 at 01:25:05PM -0400, Nicholas Berry wrote: Hi. The sense data says a SCSI device/bus or power on reset occurred. All is not well. I assume you're using Linux, from the /dev/sg1 reference. Get the mtx package installed, and see what that makes of your changer. Nik Jason Brooks [EMAIL PROTECTED] 06/20/02 06:36PM Hello, How do I get amanda to use the barcode reader on my brandnew scsi-changer? (it's an independant changer with its own scsi-generic interface) the changer appears to work otherwise, though I haven't actually begun testing backups on it... the ./docs/TAPE.CHANGER file contains the following: To initialize the label mapping you can also do an amtape config show. This will put the mapping for all tapes in the magazine into the mapping file. This doesn't work for me. enclosed is my changer.conf file and the relevent entries from my amanda.conf file. I would like to have amanda cross-reference barcodes to tape labels, perhaps with an inventory control to verify barcode/amlabel mapping. if amand needs a certain label, it tells the chager to load barcode #n. //cut from amanda.conf tpchanger chg-scsi tapedev 0 # the abstract changer config in the changer file changerfile /usr/local/etc/amanda/changer/changer.conf //my changer.conf contents: number_configs 1 eject 0 sleep 140 cleanmax 20 changerdev /dev/sg1 config 0 drivenum0 dev /dev/nst1 startuse0 havebarcode 1 enduse 12 cleancart 13 statfile/var/amanda/DailySet2/tape0-slot cleanfile /var/amanda/DailySet2/tape0-clean usagecount /var/amanda/DailySet2/totaltime tapestatus /var/amanda/DailySet2/tapestatus labelfile /usr/local/etc/amanda/changer/labelfile //sample contents of a chg-scsi*debug file chg-scsi: debug 1 pid 1218 ruid 11600 euid 11600 start time Wed Jun 19 15:27:48 2002 chg-scsi: $Id: chg-scsi.c,v 1.1 2002/06/18 18:25:41 jason0 Exp $ ARG [0] : /usr/local/libexec/chg-scsi ARG [1] : -slot ARG [2] : current Number of configurations: 1 Tapes need eject: No Inv. auto update: No barcode reader : Yes Tapes need sleep: 140 seconds Cleancycles : 20 Changerdevice : /dev/sg1 Labelfile : /usr/local/etc/amanda/changer/labelfile Tapeconfig Nr: 0 Drivenumber : 0 Startslot : 0 Endslot : 12 Cleanslot : 13 Devicename: /dev/nst1 changerident : none SCSITapedev : none tapeident : none statfile : /var/amanda/DailySet2/tapestatus Slotfile : /var/amanda/DailySet2/tape0-slot Cleanfile : /var/amanda/DailySet2/tape0-clean Usagecount: /var/amanda/DailySet2/totaltime # START SenseHandler
Amanda, clearcase and Veritas
Hello, I am about to build a wrapper script as a stand-in for vxdump. this wrapper will a) lock all of the clearcase vobs b) mount a snapshot volume c) unlock the vobs d) run vxdump on that snapshot volume e) unmount the snapshot volume. Has anyone already done this, and what sorts of problems have you had? thanks... --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
use of changer
Hello, How do I get amanda to use the barcode reader on my brandnew scsi-changer? (it's an independant changer with its own scsi-generic interface) the changer appears to work otherwise, though I haven't actually begun testing backups on it... the ./docs/TAPE.CHANGER file contains the following: To initialize the label mapping you can also do an amtape config show. This will put the mapping for all tapes in the magazine into the mapping file. This doesn't work for me. enclosed is my changer.conf file and the relevent entries from my amanda.conf file. I would like to have amanda cross-reference barcodes to tape labels, perhaps with an inventory control to verify barcode/amlabel mapping. if amand needs a certain label, it tells the chager to load barcode #n. //cut from amanda.conf tpchanger chg-scsi tapedev 0 # the abstract changer config in the changer file changerfile /usr/local/etc/amanda/changer/changer.conf //my changer.conf contents: number_configs 1 eject 0 sleep 140 cleanmax 20 changerdev /dev/sg1 config 0 drivenum0 dev /dev/nst1 startuse0 havebarcode 1 enduse 12 cleancart 13 statfile/var/amanda/DailySet2/tape0-slot cleanfile /var/amanda/DailySet2/tape0-clean usagecount /var/amanda/DailySet2/totaltime tapestatus /var/amanda/DailySet2/tapestatus labelfile /usr/local/etc/amanda/changer/labelfile //sample contents of a chg-scsi*debug file chg-scsi: debug 1 pid 1218 ruid 11600 euid 11600 start time Wed Jun 19 15:27:48 2002 chg-scsi: $Id: chg-scsi.c,v 1.1 2002/06/18 18:25:41 jason0 Exp $ ARG [0] : /usr/local/libexec/chg-scsi ARG [1] : -slot ARG [2] : current Number of configurations: 1 Tapes need eject: No Inv. auto update: No barcode reader : Yes Tapes need sleep: 140 seconds Cleancycles : 20 Changerdevice : /dev/sg1 Labelfile : /usr/local/etc/amanda/changer/labelfile Tapeconfig Nr: 0 Drivenumber : 0 Startslot : 0 Endslot : 12 Cleanslot : 13 Devicename: /dev/nst1 changerident : none SCSITapedev : none tapeident : none statfile : /var/amanda/DailySet2/tapestatus Slotfile : /var/amanda/DailySet2/tape0-slot Cleanfile : /var/amanda/DailySet2/tape0-clean Usagecount: /var/amanda/DailySet2/totaltime # START SenseHandler Ident = [L500 632], function = [generic] # START GenericSenseHandler # START DecodeSense GenericSenseHandler : Sense Keys Extended Sense ASC 29 ASCQ 00 Sense key 06 Unit Attention Sense2Action START : type(8), ignsense(0), sense(06), asc(29), ascq(00) Sense2Action generic start : Sense2Action generic END : no match for generic return - -1/Default for SENSE _UNIT_ATTENTION # STOP GenericSenseHandler STOP SenseHandler -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Variables to control tape changer
Hello, I have just acquirred a Quantum/ATL L500 library. -yay- I believe I have it functioning correctly, but there sure appears to be some arcane information regarding the changerfile configuration for chg-scsi. Here are my thougts/questions: 1) why would I want to use scsitapedev instead of dev in my changer config file? 2) What is the purpose of the sleep variable? Is it the time amanda will wait from when the changer returns to when the tape can be used? I set mine to 140 seconds as the documentation for my tape drive shows an average of 133 seconds to load a blank tape. 3) it is not clear if I can tell amanda to tell the tape changer a barcode value to load instead of searching the slots... = Now here are the specifics My amanda server is a dual pIII redhat 7.0 linux box, running amanda-2.4.3b3. this machine has the following tape hardware attached: /dev/nst0: sony dds3 4mm drive only (currently used to do backups) /dev/nst1: quantum/atl dlt8000 (tape on library) /dev/sg1: Tape changer mechanism. #dev= scsitapedev=/dev/nst1 amcheck-server: slot n: tape_rdlabel: tape open: 0: no such file or directory my chg-scsi file showed: warning open of 0: failed #dev= scsitapedev=/dev/sg0 I can't recall, but it didn't work dev=/dev/nst1 #scsitapedev= This is where I got the thing working correctly. perhaps /dev/sg0 would have worked, if I had unloaded kernel module st, and reloaded sg. I didn't want to do that though since I am running backups on /dev/nst0. What do you all think? Should I also post this to amanda-developers? --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Recommendations sought for DLT drive
Hello, I am in the market to buy a DLT 7000 library. It would be awesome if you guys could make a recommendation to me for a brand and model. I need it to run under: redhat linux 6.2-7.2 -x86 architecture -scsi connector I would like it to run under: Solaris 2.5.1, 7, and 8. I am also especially interested in those of you who have had BAD experiences with DLT libraries. Thanks in advance... --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: Estimates taking a long time.
Wow! I was just going to write about this. I have just turned on gnutar backups. I noticed that the estimates are basically tar operations with all of the ouput going to /dev/null. That is what's taking the time. Not to get the file's size, but to move the date through that system pipe into /dev/null. Might a ball-park reliable estimate of filesystem size be aquired with using find with the ls option? a shell script could interpret the excludes file into what find needs (-prune), capture the size field with cut, or awk then add them with eval for a sum total? I think I would be willing to help with this, if it's a viable thing and all... I expect there would be idiosynchrosies (SP?) between the estimates of tar and the estimates of find, but find would probably be faster... what do you all think? --jason On Fri, Jul 20, 2001 at 09:36:35PM +0100, Colin Smith wrote: I'm running backups on 3 Linux systems, one of the systems is a Cobalt Qube. All the backups are done using GNU tar. It works OK but the estimation time on the backups is nasty. I think I'll turn off the estimation and just run full dumps every day. The Qube is the slow system with the problem being related to the Linux filesystem code which reads directory entries; The system is a news server and has a couple of directories which have a lot of files. My question is this. Why run a separate estimate at all? Why not just monitor the last couple of backups and extrapolate? i.e. Day 1 incremental /home 100mb Day 2 incremental /home 110mb Day 3 incremental /home 120mb Day 4 incremental /home ???mb -- Colin Smith -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
amcheck bug(?)
Hello, amcheck is reporting an incorrect number of hosts checked for windows backups, though the backups themselves operate correctly. I am running amanda 2.4.2p1 on a linux redhat 7.0 system (streak). In addition to unix, I am also backing up windows servers. I have a test configuration called Test which I use to verify my client setups before I add them to the main system. My problem isn't whether backups succeed or fail. It is what amcheck -c Test reports. Currently, in my test disklist config, I have two entries for two hosts: # trask (Gabi's pc) streak //trask/d$ windows # pond (core server) streak //pond/c$ windows windows is the backup type and is defined in my amanda.conf file. Amcheck only reports: Amanda Backup Client Hosts Check Client check: 1 host checked in 1.482 seconds, 0 problems found (brought to you by Amanda 2.4.2p1) Why? or does it only check the client streak as shown in the disklist? if amcheck only shows the machine that smbclient is run on, should there be additional info showing that the actual connection to the windows hosts succeeeded? again, the backups themselves are running just fine. --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: upgrading from DAT to DLT
Hello, i'm sorry as this is definitely a newbie question, however, i don't have much experience with backup devices. I understand. I recommend you pick up the book unix backup and restore. this book also has a HUGE section on amanda. This is the same section online whose link can be found at http://www.amanda.org. i was recently appointed as UNIX admin here (no previous admins) and i began using my workstation with a 20/40 GB DAT drive as an amanda server to do simple network backups for a few of the mission-critical boxes (just to get something going). S, Thinking about giving that dat drive away? I probably can manage with one... :) however, now i'd like to get serious and plug in the PowerStor L200 drive (8 DLT Tapes) someone here bought over a year ago that's just been sitting around and start doing some real backup stuff. my question is, however, what would be the device name for the multi-tape changer under linux. will it remain /dev/[n]st0 or will it change to something else? Seriously, it depends. I assume both the new and old drives are scsi. if you plug in the new drive onto the scsi chain with the old one, /dev/[n]st0 will point to the drive with the lowest scsi id number. /dev/[n]st1 will point the higher scsi device id. If you replace the old drive entirely with the new one, (same id perhaps), then /dev/[n]st0 will point for certain at the new drive. the bootup messages, messages logfile, and dmesg will show you what device the kernel has assigned to your tape drives. Here's the output from my dmesg command: st: bufsize 32768, wrt 30720, max buffers 5, s/g segs 16. Detected scsi tape st0 at scsi0, channel 0, id 4, lun 0 I would be tempted to put the new drive on the chain with the old one and set its scsi device id higher than the old one. this way you can still run backups on the old drive without changing your current configuration. After that, build an entirely new amanda configuration for the new drive. Remember the value of if it ain't broke, don't fix it. again, sorry this is so basic. This is not basic. it is a relatively painless process, however. After you install the new drive, spend some time testing it, including scripts to change to different tapes, what device settings turn off and on compression etc. does anyone else have an L200 that would possibly be willing to post their chg-multi settings (or not)? the tapetype settings for your particular tape cartridges should be pretty easy to find on the tapetype list in the faq-o-matic. I do hope someone here has that tape drive type for controlling the more advanced features of your drive. thanks in advance for any assistance. You are welcome. I hope this helps... Highest Regards, Edwin R. Rivera UNIX Administrator Tel: +1 305 894 4609 Fax: +1 305 894 4799 [EMAIL PROTECTED] -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: amdump error
Hello,
First thing this morning:
The problem repeated itself. I even copied the configuration to a test configuration
and amdump ran just fine.
1500:
I just ran amdump by hand and got a good backup while at a doctor's office. I had
changed one thing: I commented out of my disk list all
entries with the following dumptype used:
define dumptype nocomp-root-incr {
nocomp-root
comment incremental only
skip-full
}
Is there anything obviously wrong about this?
2335:
Wait! the cron job just failed! Yikes!
Is there any chance the run from the night before hasn't finished?
There are no leftover processes, and amcleanup run as my amanda user shows nothing to
recover from
[streak : bv-amanda]$ amcleanup Daily
amcleanup: no unprocessed logfile to clean up.
Scanning /amanda1...
lost+found: skipping cruft directory, perhaps you should delete it.
[streak : bv-amanda]$
I am running amanda version 2.4.2p1 on both server and all clients. The server is a
redhat linux 7.0 machine running a distclean compile of
kernel 2.2.19 (not redhat's version). I have not patched amanda with anything (yet),
but the fstab issues I mentioned in a previous post,
do not exist here. (that is, I have removed the label scheme used in redhat 7.0's
/etc/fstab).
I am running linux's version of dump, although that's irrelevant here.
Was the machine possibly rebooted while the previous run was still
going and amcleanup was not run during the boot sequence?
It was rebooted several times during the day (the day after the first failure), but no
amanda sequences were being run at the time. I have
also updated my rc.local with the following script exerpt:
echo cleaning up any amanda dumps left in the lurch...
for i in `ls /usr/local/etc/amanda |grep -v ^lbl`
do
echo $i
su -l bv-amanda -c /usr/local/sbin/amcleanup ${i}
done
If the above ideas don't help, you could add a couple of lines to the
script right before it exits:
ls -l $errfile $logdir/log 12
ps -fu AMANDA 12
Here's the ouput, but it just worked when I ran amdump by hand as my amanda user. the
cron entry is in my amanda user's crontab...
start jason's debug
ls: /var/amanda/DailySet1/amdump: No such file or directory
ls: /var/amanda/DailySet1/log: No such file or directory
UIDPID PPID C STIME TTY TIME CMD
1160021886 21885 0 Apr30 pts/000:00:00 -bash
11600 2050 2049 0 11:36 pts/300:00:00 -bash
11600 2805 2050 0 11:46 pts/300:00:00 sh /usr/local/sbin/amdump Daily
11600 2810 2805 0 11:46 pts/300:00:00 ps -fu bv-amanda
I guess I will see what happens tonight...
--jason
--
~~~
Jason Brooks ~ (503) 641-3440 x1861
Direct ~ (503) 924-1861
System / Network Administrator
Wind River Systems
8905 SW Nimbus ~ Suite 255
Beaverton, Or 97008
partition name conversions
Hello, With the advfs patch available, I can now use directory names instead of partition names. Is there a way to convert to directory names from partition names in the middle of a dumpcycle and have amanda know they are the same filesystems? --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Test dumps
Hello, Is there a way to run amdump on my current configuration without it counting against me? I think I need to do: 1) set tapedev to /dev/null 2) changerfile is already commented out 3) index=no Is there more? --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
amdump error
Hello, I am trying to post my messages atomically: one topic at a time. Please let me know if this is not correct... I am trying to see why my cron job amdump Daily is sometimes failing with the message: amdump: amdump or amflush is already running, or you must run amcleanup This configuration has functioned before. Cron is executing as my amanda user, and all permissions (I believe) are correct. Running amcleanup Daily produces amcleanup: no unprocessed logfile to clean up. I looked in the amdump script to find the line amdump or amflush is already running: if test -f $errfile || test -f $logdir/log; then echo amdump: amdump or amflush is already running, or you must run amcleanup 2 exit 1 fi Well, $errfile corresponds to /var/amanda/DailySet1/amdump and $logdir/log corresponds to /var/amanda/DailySet1/log. Neither one of these files exist, so why is amdump balking at this point? --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
increasing level zero dumps per tape cycle (?)
Hello, I am suffering a bit from terminology shock. Here's what I want to know: I have thirty tapes in my tape cycle. I want to be certain to get TWO level zero dumps of each file system in this tape cycle. Is this as simple as having my dumpcycle equal to 15? (leaving out extra tapes for the moment...) --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
amanda applications questions
Hello, can I run the planner program independantly to see how amanda is scheduling dumps? I realize planner puts out estimations based on my tape type info, netusage limits, etc. --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Disklist formatting questions:
Hello, I just tried doing a test dump and ran into two problems: 1) When I changed the partitions in my disklist file to the mount points, it worked except for /. I did this so the logs and indexes made sense without having to cross-reference each fstab. example: jetstream hda8 nocomp-root jetstream hda1 nocomp-root become jetstream / nocomp-root jetstream /boot nocomp-root the error messages complain about not being able to open /dev/root. Is there a mnemonic to indicate to amanda backup the root filesyste aside from just the device file? 2) Redhat 7.0 changes its fstab usage to use the labeling feature of the ext2 filesystem. example: LABEL=/ /ext2defaults 1 1 LABEL=/boot /bootext2defaults 1 2 Amanda would chock on these lables until I redid the fstab files to read: /dev/hda6/ext2defaults 1 1 /dev/hda1/bootext2defaults 1 2 Does amanda have the ability to interpret the LABEL= portion of fstab then dig the information out of ext2fs? If not, is it planned? I am ok with changing the fstab file, but it kind of falls into my philosophy of if it ain't broke, don't fix it. --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
Re: Tar, general, and amindexd questions
Hello,
I forgot to include I am running version 2.4.2p1 on all systems.
I am running amrecover on the amanda server to remove any networking
hardware issues.
On Thu, Apr 26, 2001 at 02:30:23PM -0700, Jason Brooks wrote:
Hello,
Thank you for responding to my message. this platform (redhat 7.0) uses
Xinetd. here's the configuration:
[streak : xinetd.d]$ cat amandaidx
# default: on
# this is for the amindexd program.
service amandaidx
{
socket_type = stream
protocol= tcp
wait= no
user= bv-amanda
groups = yes
server = /usr/local/libexec/amindexd
server_args = amindexd
}
Note that amindexd in the path. That says your inetd.conf or xinetd
config file is wrong. It's passing the program name (amindexd) as an
arg to the program, who is taking it as the config name to use.
Ok, I see (I think) how having my configuration as part of the
server_args line could have an effect, but what if I wish to
run amindexd for a different configuration such as Test, Periodic_full,
or Daily? Since I specify amrecover -C Daily on the command line, I
would expect the data stream from client to server to tell amindexd
what configuration to look for.
Thanks for your time...
--jason
--
~~~
Jason Brooks ~ (503) 641-3440 x1861
Direct ~ (503) 924-1861
System / Network Administrator
Wind River Systems
8905 SW Nimbus ~ Suite 255
Beaverton, Or 97008
--
~~~
Jason Brooks ~ (503) 641-3440 x1861
Direct ~ (503) 924-1861
System / Network Administrator
Wind River Systems
8905 SW Nimbus ~ Suite 255
Beaverton, Or 97008
Re: Tar, general, and amindexd questions
Hello, On Thu, Apr 26, 2001 at 05:53:13PM -0500, John R. Jackson wrote: server_args = amindexd Ditch that line. None of the Amanda xinetd config files should have a server_args key/value. That did it. it is working now. BTW, the -C on amrecover is optional and probably should not be there. Do you mean to say that the -C option shouldn't be optional? --jason -- ~~~ Jason Brooks ~ (503) 641-3440 x1861 Direct ~ (503) 924-1861 System / Network Administrator Wind River Systems 8905 SW Nimbus ~ Suite 255 Beaverton, Or 97008
