Increase parallel processes
Hi All, I have 5 main server to archive. The largest one is about 6TB (RAID 5) containing 70% of my DLEs both in numbers and data volume. Amanda does a good job of running parallel dumps at the start of each run, however after a few hours she is left with only DLE to process on the RAID 5 server. I am now running out of quite hours in which to do the nightly archiving. Any suggestions on getting Amanda to parallel the dumps from the RAID 5 server? Thanks, Vytas -- Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Ext. 103 Fax: (613) 725-1716 mailto:vy...@dendron.com
Re: Client backup fails with Timeout error
Hi All, Up to about a week ago all my clients (Linux) have been archiving properly. For the past 5 runs of Amanda one client is not backing up. Below are the error message I am getting. Searching the list archives did not point me to anything that may be related to my environment. Amanda server and clients are running Debian packages. On the server I am running Amanda-2.5.1p3 and on all clients Amanda-2.5.1p1. I have re-booted the server and the client between runs with no success. As far as I can tell no changes were made to the troublesome client which may account for the failures. All computer are inside the firewall. Any ideas of what may be going wrong? Amstatus Message hal2:/mnt/hdb1/Can_Fin_photos 1 driver: (aborted:[request failed: timeout waiting for REP])(too many dumper retry) Amanda Report hal2/mnt/hdb1/Can_Fin_photos lev 1 FAILED [cannot read header: got 0 instead of 32768] Thanks, Vytas -- Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Ext. 103 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
One DLE too many!
Hi Again,
It's taken me a couple of days to get the suggestions tested.
So this is my current DLE entries and output from the amanda report
and comments from amrecover: Why are the DLEs with the
includes failing? In the reports they claim to be succeeding but
they are not doing what I want. Any attempt to recover the
files fails using amrecover, which I am not surprised about . Do I
need to add a wildcard (*) to the include directory names? Or am I
missing something else?
Thanks Again!
Vytas
hal /mnt/data06/Deforest3_PART1 /mnt/data06/Deforest3 {
comp-user-tar
include file optional ./AirPhoto_Load_Tool
include file optional append ./Nova_Scotia_Photos
} 87
planner: Forcing full dump of hal:/mnt/data06/Deforest3_PART1 as directed.
hal/mnt/data06/Deforest3_PART1 010
1 10.0 0:02 0.0 0:00 2.3
Report states that it backed up 10KB as Level 0. The two
directories contains over 40 GB of data
amrecover setdisk /mnt/data06/Deforest3_PART1
200 Disk set to /mnt/data06/Deforest3_PART1.
No index records for disk for specified date
If date correct, notify system administrator
/
hal /mnt/data06/Deforest3_PART2 /mnt/data06/Deforest3 {
comp-user-tar
include file optional ./delivery*
include file optional append ./Imagery
} 87
planner: Forcing full dump of hal:/mnt/data06/Deforest3_PART2 as directed.
hal/mnt/data06/Deforest3_PART2 010
1 10.0 0:02 0.0 0:00 2.3
Report states that it backed up 10KB as Level 0. The two
directories contains over 8 GB of data
amrecover setdisk /mnt/data06/Deforest3_PART2
200 Disk set to /mnt/data06/Deforest3_PART2.
No index records for disk for specified date
If date correct, notify system administrator
/
hal /mnt/data06/Deforest3_REST /mnt/data06/Deforest3 {
comp-user-tar
exclude file optional./AirPhoto_Load_Tool
exclude file optional append ./Nova_Scotia_Photos
exclude file optional append ./delivery
exclude file optional append ./Imagery
} 87
planner: Forcing full dump of hal:/mnt/data06/Deforest3_REST as directed.
hal/mnt/data06/Deforest3_REST 0 30640990
22670888 74.0 114:19 3305.2 FAILED --
Backup file flushed perfectly and appears to contain all the required
excluded directories.
amrecover setdisk /mnt/data06/Deforest3_REST
200 Disk set to /mnt/data06/Deforest3_REST.
amrecover ls
2006-03-16 .
2006-03-16 NAPL/
2006-03-16 NS_Oblique_Aerials/
2006-03-16 Quality_Control/
2006-03-16 Received/
2006-03-16 readme.txt
Vytas Janusauskas
Dendron Resource Surveys Inc.
880 Lady Ellen Place, Suite 206
Ottawa, Ontario, Canada
K1Z 5L9
Voice: (613) 725-2971
Fax: (613) 725-1716
mailto:[EMAIL PROTECTED]
Re: One DLE too many!
Hi Paul,
This e-mail thread has produce quite a bit of response. All
appreciated. However I am still not clear if I can split my DLE's
using the include and exclude commands. I am under the impression
that amdump runs the backup as root. Should it therefore not have
access to all directories? I have added the information you
requested inline below.
Thanks for taking the time to steer me in the right direction.
Vytas
less /essAt 09:29 AM 3/17/06, Paul Bijnens wrote:
On 2006-03-17 14:53, Vytas Janusauskas wrote:
Hi Again,
It's taken me a couple of days to get the suggestions tested.
So this is my current DLE entries and output from the amanda report
and comments from amrecover: Why are the DLEs with the includes
failing? In the reports they claim to be succeeding but they are
not doing what I want. Any attempt to recover the files fails
using amrecover, which I am not surprised about . Do I need to
add a wildcard (*) to the include directory names? Or am I
missing something else?
What are the permissions of the top level directory /mnt/data06/Deforest3 ?
Listed below are the directory permissions listed from /mnt and up.
drwxr-xr-x 15 root root 4096 Nov 2 11:37 mnt
drwxrwsr-x 8 root uucp 4096 Mar 3 10:26 data06
drwxrws--- 10 root uucp 4096 Mar 16 09:32 Deforest3
drwxrws--- 3 root uucp 4096 Jan 10 15:23 AirPhoto_Load_Tool
drwx--S--- 5 root uucp 4096 Mar 1 11:18 Imagery
drwxrws--- 2 root uucp 20480 Feb 23 11:09 NAPL
drwxrws--- 4 root uucp 4096 Jan 16 07:48 NS_Oblique_Aerials
drwxrws--- 2 root uucp 24576 Mar 6 16:07 Nova_Scotia_Photos
drwxrws--- 2 root uucp 4096 Mar 9 09:55 Quality_Control
drwxrws--- 2 root uucp 4096 Mar 6 13:48 Received
drwxrws--- 11 root uucp 4096 Mar 7 09:11 delivery
That directory needs to be readable and searcheable by the amanda user.
(and the all the parent directories, up to the root, too!)
I have a directory on /mnt/data06/Deforest4 which does not use any
include or exclude which appears to archive properly.
hal /mnt/data06/Deforest3_PART1 /mnt/data06/Deforest3 {
comp-user-tar
include file optional ./AirPhoto_Load_Tool
include file optional append ./Nova_Scotia_Photos
} 87
By putting the word optional here, you are in fact suppressing any
error messages about this file.
But in this case you do not want that at all.
Usually you do not want that!!!
So omit it.
I put the word optional in to suppress the amcheck error messages. I
did not believe it was also used by amdump.
Backup file flushed perfectly and appears to contain all the
required excluded directories.
I guess you mean something like the backup file does not contain any of
the excluded directories as required ?
That is correct.
amrecover setdisk /mnt/data06/Deforest3_REST
200 Disk set to /mnt/data06/Deforest3_REST.
amrecover ls
2006-03-16 .
2006-03-16 NAPL/
2006-03-16 NS_Oblique_Aerials/
2006-03-16 Quality_Control/
2006-03-16 Received/
2006-03-16 readme.txt
Vytas Janusauskas
Dendron Resource Surveys Inc.
880 Lady Ellen Place, Suite 206
Ottawa, Ontario, Canada
K1Z 5L9
Voice: (613) 725-2971
Fax: (613) 725-1716
mailto:[EMAIL PROTECTED]
Re: One DLE too many!
Hi Paul,
Listed below is what I get when I try to define a DLE with the
include statement. However last night I ran it with the include line
commented out and it backed up the complete Deforest3 data set as
expected . However as soon as I add
include ./AirPhoto_Load_Tool to the disklist I get amcheck
errors about Can't open disk and No include for
I searched the list and docs and cannot see what I am doing wrong.
Thanks again,
Vytas
My test disklist entry is:
hal /mnt/data06/Deforest3_PART1 /mnt/data06/Deforest3 {
comp-user-tar
include ./AirPhoto_Load_Tool
} 87
Dumptype is:
define dumptype comp-user-tar {
program GNUTAR
comment partitions dumped with tar
options compress-fast, index
comprate 0.75, 0.75
priority high
}
amadmin output is:
host hal:
interface default
disk /mnt/data06/Deforest3_PART1:
device /mnt/data06/Deforest3
program GNUTAR
include file ./AirPhoto_Load_Tool
priority 2
dumpcycle 30
maxdumps 8
maxpromoteday 1
strategy STANDARD
compress CLIENT FAST
comprate 0.75 0.75
auth BSD
kencrypt NO
holdingdisk YES
record YES
index YES
skip-incr NO
skip-full NO
Test directory listing is:
hal:/mnt/data06# du -h --max-depth=1 Deforest3
248KDeforest3/AirPhoto_Load_Tool
2.7GDeforest3/delivery
1.8GDeforest3/NS_Oblique_Aerials
6.4GDeforest3/Imagery
17M Deforest3/Received
1.6MDeforest3/Quality_Control
40G Deforest3/Nova_Scotia_Photos
28G Deforest3/NAPL
78G Deforest3
amcheck output is:
vxa2:/etc/amanda/Den$ /usr/sbin/amcheck -c
Den
Amanda Backup Client Hosts Check
ERROR: hal: [Can't open disk '/mnt/data06/Deforest3']
ERROR: hal: [No include for '/mnt/data06/Deforest3_PART1']
Client check: 5 hosts checked in 3.259 seconds, 2 problems found
(brought to you by Amanda 2.4.4p3)
At 02:37 PM 3/10/06, Paul Bijnens wrote:
Vytas Janusauskas schreef:
snip ...
You can use inline dumptypes (with
brace-newline-dumptypedefinition-newline-endbrace),
and explicitly make a the DLE-name different from the directory to
differentiate the different includes:
client1.example.com /data06/Deforest3_PART1 /data06/Deforest3 {
comp-user-tar
include ./Air*
include append ./Nova?Scotia?Photos
} 1
Vytas Janusauskas
Dendron Resource Surveys Inc.
880 Lady Ellen Place, Suite 206
Ottawa, Ontario, Canada
K1Z 5L9
Voice: (613) 725-2971
Fax: (613) 725-1716
mailto:[EMAIL PROTECTED]
Re:One DLE too many!
Hi All, I have been using Amanda (current version 2.4.4p3) for the last couple of years without any major issues. However recently due to an influx of new projects the volume of data has grown. What is happening is that users are adding new directories to the data servers and not notified me. So all of a sudden some of my DLEs which have been merrily archiving have started to exceed my available tape capacity. I have toyed with the idea of using tar excludes to address this issue, unless I am missing something would I have to create a separate DumpType for each DLE? For better illustrate the issue for the following DLE they added the 40G Nova Scotia Photos directory. I was planning on breaking up the DLE listed below into 8 individual entries but that may lead to problems of handling the directories with whitespaces in their names. If I was just dealing with a few DLEs separating them into smaller chunks would not be an issue, however I have already fragmented my data into 240 + DLEs. hal /mnt/data06/Deforest3 comp-user-tar 87 hal:/mnt/data06# du -h --max-depth=1 Deforest3 248KDeforest3/AirPhoto Load Tool 40G Deforest3/Nova Scotia Photos 2.7GDeforest3/delivery 1.8GDeforest3/NS_Oblique_Aerials 6.4GDeforest3/Imagery 17M Deforest3/Received 1.6MDeforest3/Quality Control 28G Deforest3/NAPL 78G Deforest3 Any suggestion of how best to deal with this situation is greatly appreciated. Vytas Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re: Restoring without Amanda
At 02:13 PM 9/7/05, you wrote: On Wed, Sep 07, 2005 at 10:32:47AM -0400, Joshua Baker-LePain wrote: snip ... Vytas, Tony, Would the header message have been any clearer if it said dd if=tape bs=32k skip=1 | /bin/gzip -dc | bin/tar -...f - or maybe dd if=tape bs=32k skip=1 | /bin/gzip -dc | bin/tar -tar_options -f - That would help make thing clearer. May I also suggest that something similar to what Paul Bijnens posted be added to the docs. I appreciate the fast response I got on my query from Paul ,Tony and yourself. I usually extract with these options: $ su # we need to be root to be able to restore owner, etc. # mkdir /the/lost/directory # where to place the files # cd /the/lost/directory # dd if= bs=32k skip=1 | /path/to/gnutar -xzpvf - -x : extract -z : filter through gzip -p : preserve permissions -v : some animation on the screen -f - : read from standard input Vytas snip ... Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re: Restoring without Amanda
Hi All, I had a failure of my Amanda index drive. I am now trying to restore some archives. What I have done so far is the following: Ran amrestore to get a listing of tape contents dendron:/mnt/dump_drive/restore# mt -t /dev/nst0 rewind dendron:/mnt/dump_drive/restore# amrestore /dev/nst0 no.such.host amrestore: 0: skipping start of tape: date 20050304 label Project223 amrestore: 1: skipping dendron.l8r.net._mnt_vxa2_scans_c_Seven__Pent22.20050304.0 amrestore: 2: skipping dendron.l8r.net._mnt_vxa2_scans_c_Seven__Unix.20050304.0 amrestore: 3: skipping dendron.l8r.net._mnt_vxa2_scans_c_Seven__Pent21.20050304.0 amrestore: 4: skipping dendron.l8r.net._mnt_vxa2_scans_c_Seven__Final__Sids.20050304.0 amrestore: could not fsf /dev/nst0: Input/output error determined that what I need is in file 2: dendron.l8r.net._mnt_vxa2_scans_c_Seven__Unix.20050304.0 So I: mt -t /dev/nst0 rewind mt -t /dev/nst0 fsf 2 dd if=/dev/nst0 bs=32k count=1 This list the following: AMANDA: FILE 20050304 dendron.l8r.net /mnt/vxa2/scans/c/Seven_Unix lev 0 comp .gz program /bin/tar To restore, position tape at start of file and run: dd if=tape bs=32k skip=1 | /bin/gzip -dc | bin/tar -f... - So I rewind the tape skip 2 files and then issue: dd if=/dev/nst0 bs=32k skip=1 | /bin/gzip -dc | /bin/gtar -f I tried various things as substitutes for such as /mnt/vxa2/scans/c/Seven_Unix and dendron.l8r.net._mnt_vxa2_scans_c_Seven__Pent22.20050304.0 Results is always an error similar to the following: /bin/tar: dendron.l8r.net._mnt_vxa2_scans_c_Seven__Unix.20050304.0: Cannot open: No such file or directory /bin/tar: Error is not recoverable: exiting now /bin/tar: Child returned status 2 /bin/tar: Error exit delayed from previous errors 1+0 records in 0+0 records out 0 bytes transferred in 0.021878 seconds (0 bytes/sec) Any pointer to what I should be putting on the final command line would be appreciated. Thanks, Vytas Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Samba and local mounts timeout
Amanda has been purring along for months now, however the last 4 days she has been timing out on all drives Samba and local server mounted drivers. As far as I know nothing has been changed except that regular updates are done to the Debian distribution using apt-get. System is running Linux 2.4.23-1-686 Samba version is 3.0.2a-Debian Amanda version is 2.4.4p2 Amanda config has dtimeout of 1800 and etimeout of 300. I get the following error messages in the log file for each Samba mount FAIL planner vxa2 /data03/archive/pent21/e 20040318 0 [Estimate timeout from vxa2] Also all local drives are timing out with the following message FAIL planner vxa2 /scans/d 20040318 0 [Estimate timeout from vxa2] The strange thing is that it continues to backup another Linux servers files. Anyone have any ideas of what has got Amanda all upset? Thanks, Vytas Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re: Missing files from Samba backup
Hi Amanda Users! This morning I noticed that a DLE which has been forced (Level 0) on a NT 4.0 Service Pack 6 computer via Samba did not completely backup all files. It seem that for any file/directory listed in Amanda's report (See below) which start with a ? SUCCESS - 0 statement were not archived or added to the index files. The reason I found that there was a serious problem is that a full should contain about 88 gigs of data while only about 20 gigs were taped. This caused me to look back through the full backup logs for other computers and found similar problems. At first I thought it was due to the NT box drive being a RAID configuration be alas the problem also occurs on Win 85 and Win 98 computers. A search of the archive did not point to any solutions. Anyone have an idea of what is going wrong? Amanda version 2.4.2p2 tar (GNU tar) 1.13.25 Thanks, Vytas These dumps were to tape Daily14. The next tape Amanda expects to use is: Daily15. FAILURE AND STRANGE DUMP SUMMARY vxa2 //nt-01/nt01-d lev 0 STRANGE /-- vxa2 //nt-01/nt01-d lev 0 STRANGE sendbackup: start [vxa2://nt-01/nt01-d level 0] sendbackup: info BACKUP=/usr/bin/smbclient sendbackup: info RECOVER_CMD=/usr/bin/smbclient -f... - sendbackup: info end ? NT_STATUS_SHARING_VIOLATION opening remote file \pagefile.sys (\) ? SUCCESS - 0 opening remote file \Project\AgCan\0_Ontario\ALCDB_Vytas.zip (\Project\AgCan\0_Ontario\) ? SUCCESS - 0 opening remote file \Project\AgCan\0_Ontario\Arc_Info_Generalize.bmp (\Project\AgCan\0_Ontario\) ? SUCCESS - 0 listing \Project\AgCan\0_Ontario\bend100\* ? SUCCESS - 0 listing \Project\AgCan\0_Ontario\bend15\* ? SUCCESS - 0 listing \Project\AgCan\0_Ontario\bend30\* ? SUCCESS - 0 listing \Project\AgCan\0_Ontario\bend300\* ? SUCCESS - 0 listing \Project\AgCan\0_Ontario\bend50\* ? SUCCESS - 0 listing \Project\AgCan\0_Ontario\test\* ? SUCCESS - 0 opening remote file \Project\AgCan\0_Ontario\Testdata.mxd ... ... ... ? SUCCESS - 0 listing \Proposal\* ? SUCCESS - 0 listing \restore\* ? SUCCESS - 0 opening remote file \ROND.DBF (\) ? SUCCESS - 0 listing \scan2cad\* ? SUCCESS - 0 listing \Scan2Vec\* ? SUCCESS - 0 listing \Seagate Software\* ? SUCCESS - 0 opening remote file \SQL.LOG (\) ? SUCCESS - 0 listing \System\* ? SUCCESS - 0 listing \TCWIN\* ? SUCCESS - 0 listing \wag\* ? SUCCESS - 0 listing \wagner_unix\* ? SUCCESS - 0 listing \WINNT\* ? SUCCESS - 0 listing \WorkSpace\* ? SUCCESS - 0 listing \yukon scans\* | tar: dumped 29441 files and directories | Total bytes written: 21413944832 sendbackup: size 20912056 sendbackup: end \ DUMP SUMMARY: DUMPER STATS TAPER STATS HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s -- -- vxa2 -nt-01/nt01-d 0 20912056 14609312 69.9 126:22 1926.7 39:11 6214.0 Window NT drive stats. Total Files Listed: 184354 File(s) 88,102,672,147 bytes 29,309,644,800 bytes free Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re: Amrecovering files/directories with blanks.
Hi Again, Some of my windows users are naming directories and files with space characters as separators between words. While doing a ls on amrecover displays these space containing files/directories how do you go about changing into these directories setdisk strange data directory does not work neither does setdisk strange\ data\ directory There must be a way to do it! I would prefer not to have to restore the full DLE if I can avoid it. Thanks again, Vytas Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re: Premature overwriting of level 0
Hi, Setup: 20 tapes (VXA-2 v-23) Native capacity 67GB. 1 dump cycle. Amanda version 2.4.2p2 Holding disk 105GB All DLE are Samba shares for Win95/98/NT6/Win2000. Problem: I am now 7 tapes into the second rotation and the level 0 backups are being overwritten. It seems to be due to the large size of incrementals being archived on a daily basis. There are numerous days where Amanda only does incremental (level 1) and even so Amanda is not able to fit them all onto the daily tape. Paul (see below) suggested as one of the options is to patch Amanda to allow a higher level (2,3,4...) backups to be done with Samba shares, thereby reducing the size of daily incrementals. I know that this would result in greater numbers of tape that would have to be accessed to restore the data but I am willing to live with that issue. Solutions? From Paul's response I may have wrongly assumed that all that would be required to fake level 2, 3 etc... backups would be to toggle the windows archive bit when Amanda is doing incremental archives. Unfortunately, I am not a programmer. Though I did searched through the source code I could not find where to add the switch archive bit code. If I am not completely out to lunch, I would appreciate any pointers required to implement. If I am bouncing off the walls any other suggestions to solving my problem are appreciated as long as it does not involve additional hardware. Thanks all! Vytas At 11:07 AM 3/24/03 +0100, Vytas Janusauskas wrote: What must I do to ensure that all my level 0 are not overwritten before they are archived a second time? I have days where lots of new files get created. Can some explain why none of my backups are ever level 2 or higher? At 11:07 AM 3/24/03 +0100, Paul Bijnens wrote: When using smbclient to backup the builtin tar command uses the archive bit of the FAT/NTFS file system to determine incremental backup. The bit contains only a enough state to emulate a level 1 incremental backup: A full backup runs: smbclient ... -Tqca - -a means reset the archive bit An incremental backup runs: smbclient ... -Tqcg - -g means only files with archive bit set, i.e. everything changed since last full backup. It's a limitation of PC filesystems and smbclient. There once existed an idea to modify the amanda sources to reset the archive bit on incremental runs too, and let amanda do the bookkeeping on the dumplevels but I've never seen patches for this... Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re: amrecover tar error.
Hi Paul,
At 01:48 AM 3/18/03 +0100, Paul Bijnens wrote:
Op ma 17-03-2003, om 18:05 schreef Vytas Janusauskas:
I am trying to restore a Windows 95 directory to an NT drive connected via
Samba and get the following error.
/1895-102.tif
./1895-103.tif
Does this mean that it restored 1 file?
Yes , that is correct.
tar: Skipping to next header
tar: Archive contains obsolescent base-64 headers
What command did you use exactly?
amrecover Project as root
setdisk
add *
extract
quit
Testing with a re-run of the same data set restored correctly. So the
cause of the restore failure is still a mystery to me. I will keep check
the backup to ensure that they are in fact restorable.
What is the amanda headerblock suggesting as command?
Typically these are errors of a corrupted tarfile, and sometimes
simply of trying to extract a gzipped tar file without the -z option.
Tape drive being used is Eliant 820 drive with the following definition
define tapetype Eliant 820 {
comment Exabyte Eliant 820 drive
length 7 gb
filemark 48 kbytes
speed 1 mb
}
As far as I can understand the tapetype here is completely irrelevant.
I agree I just put it in for completeness.
Tar version is : tar (GNU tar) 1.13.25
Amanda version is: 2.4.2p2
Disklist entry is: backup /dosworld/pent17/e/Colorado10 comp-user-tar
So you have the share smbmounted on a Unix computer named backup.
Why not using smbclient, as builtin in Amanda? At least you
can do level 1 incrementals (or maybe even more, I didn't check
in the latest amanda sources.)
The reason for mounting the drive as opposed to using smbclient is that the
directories I am backing up are completed project which are to be removed
from the Windows boxes. Instead of creating temporary shares to these
directories I just mount them and then archive the required
directories. Once I have all the project directories archived I verify the
data via restore to temporary directories and then comment out the project
directories from the disklist. This allows me to archive the projects and
then use Amanda's index files to locate and restore the archived if ever
required. If anyone has a better procedure for archiving Windows completed
project directories I am all ears.
Vytas
Vytas Janusauskas
Dendron Resource Surveys Inc.
880 Lady Ellen Place, Suite 206
Ottawa, Ontario, Canada
K1Z 5L9
Voice: (613) 725-2971
Fax: (613) 725-1716
mailto:[EMAIL PROTECTED]
Re: Premature overwriting of level 0.
Hi Amanda Users, I have 20 tape for daily archiving. I have successfully done one pass through the tapes with Amanda. I am now back to tape 4 up to this point all Level 0 archives to be overwritten have been redone. I am coming up to tape 5 tomorrow and from looking at the amoverview table I see one to the DLE level 0 will be overwritten. Today's amdump for the DLE in question states: vxa2://pent20/pent20-c 1 [dumps way too big, must skip incremental dumps] Yes all the DLE I am archiving are Windows disks. The above message has been appearing in all my Amanda reports for various DLEs. What must I do to ensure that all my level 0 are not overwritten before they are archived a second time? I have days where lots of new files get created. Can some explain why none of my backups are ever level 2 or higher? Amanda version 2.4.2p2. Thanks Vytas Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re: amrecover tar error.
Hi,
I am trying to restore a Windows 95 directory to an NT drive connected via
Samba and get the following error.
/1895-102.tif
./1895-103.tif
tar: Skipping to next header
tar: Archive contains obsolescent base-64 headers
Tape drive being used is Eliant 820 drive with the following definition
define tapetype Eliant 820 {
comment Exabyte Eliant 820 drive
length 7 gb
filemark 48 kbytes
speed 1 mb
}
Tar version is : tar (GNU tar) 1.13.25
Amanda version is: 2.4.2p2
Disklist entry is: backup /dosworld/pent17/e/Colorado10 comp-user-tar
Any pointers are appreciated.
Thanks,
Vytas
Vytas Janusauskas
Dendron Resource Surveys Inc.
880 Lady Ellen Place, Suite 206
Ottawa, Ontario, Canada
K1Z 5L9
Voice: (613) 725-2971
Fax: (613) 725-1716
mailto:[EMAIL PROTECTED]
Re: Why is the estimates off?
Hi Folks, I have been trying to archive about 20 Window's machines (Win95/98/NT). Overall things seem to be working. Can anyone explain why the estimates and real size vary so greatly? I have checked the hosts and no new files are being created between the time the estimate is done and the backup process run. Using /var/log/amanda/DailySet1/amdump from Tue Feb 18 07:48:12 EST 2003 vxa2:/dosworld/nt-01/e/part2 057724000k writing to tape (18:08:26) vxa2:/dosworld/pent05/c/ 0 3253248k finished (10:58:27) vxa2:/dosworld/pent13/c/ 0 1666944k finished (9:43:19) SUMMARY part real estimated size size partition : 3 estimated : 3 36380995k failed : 0 0k ( 0.00%) wait for dumping: 0 0k ( 0.00%) dumping to tape : 0 0k ( 0.00%) dumping : 00k0k ( 0.00%) ( 0.00%) dumped : 3 62644192k 36380995k (172.19%) (172.19%) wait for writing: 00k0k ( 0.00%) ( 0.00%) writing to tape : 1 57724000k 32109290k (179.77%) (158.67%) failed to tape : 00k0k ( 0.00%) ( 0.00%) taped : 2 4920192k 4271705k (115.18%) ( 13.52%) 4 dumpers idle : not-idle taper writing, tapeq: 0 network free kps: 600 holding space : 49794176k ( 46.31%) dumper0 busy : 10:13:20 ( 99.99%) taper busy : 0:13:36 ( 2.22%) 0 dumpers busy : 0:00:00 ( 0.00%) 1 dumper busy : 10:13:22 (100.00%)not-idle: 7:18:54 ( 71.56%) client-constrained: 2:53:58 ( 28.36%) Thanks all, Vytas Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re: Exabyte VXA - 2
Hi Jon,
Good to hear your positive comments about the VXA-2 drive.
Vytas
At 11:04 AM 1/8/03 -0500, Jonathan S. Keim wrote:
I've been pretty happy with it in the short time I've been using it. It
had about the right stats in terms of price/GB, and so far have had no
trouble. Here's the tapetype entry I've been using:
define tapetype VXA-2 {
comment produced by tapetype program 12/09/2002
length 67593 mbytes
filemark 2621 kbytes
speed 5432 kps
}
Jon
Re: Lot of data very small tape!
Hi Folks, Still trying to configure Amanda. I have about 20 WIn 95, Win 98 and NT computers to archive. So as a starting point I have picked 3 computers which in total have drive capacity of over 530 GB of which 430 GB is used. Using Samba I have successfully mounted the drives. Since I have a tape drive with a non-compressed capacity of less than 7 GB, I am forced to use individual directories in the disklist. Unfortunately I have no control of where the users will be creating new directories. How do you ensure that all the directories are accounted for in the disklist? It is also not unusual that some of these directories to be larger than the tape capacity. I plan to deal with them after Amanda complains about them. For example when the backup burps on a directory, I will them subdivide the offending directory into sub-directory's smaller than the tape capacity. Will this work or is there a better way of dealing with these larger directories? In testing I did try to backup a directory, I though that after compression it would fit on the tape. Unfortunately due to the TIFF files the directory did not compress as well as expected. So now I have a dump in the holding disk which will not fit on the tape. Do I just delete the file from the holding disk or is there a procedure I need to run so as not corrupt Amanda's index files. Thanks, Vytas Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
Re:Large partitions (2 Gigabyte +)
Hi Folks, I am trying to backup about 20 Windows (Win95/Win98/Nt6) computers using Amanda version 2.4.2p2. I mount the Windows drives via Samba without any problem. When I try to backup a mount point ie. \Dosworld\Pent14\e which is 4.1 Gb in size. I get the following log output: START planner date 20021127 INFO planner Adding new disk backup:/dosworld/pent14/e. START driver date 20021127 START taper datestamp 20021127 label Daily01 tape 0 FINISH planner date 20021127 STATS driver startup time 64.044 FAIL dumper backup /dosworld/pent14/e 0 [data write: File too large] sendbackup: start [backup:/dosworld/pent14/e level 0] sendbackup: info BACKUP=/bin/tar sendbackup: info RECOVER_CMD=/bin/gzip -dc |/bin/tar -f... - sendbackup: info COMPRESS_SUFFIX=.gz sendbackup: info end INFO taper tape Daily01 kb 0 fm 0 [OK] FINISH driver date 20021127 time 7526.614 Looking at the Google listings it would seem to fail due to the holding disk which has 26 Gigabytes of free space does not support images greater than 2 Gigabytes in size. The holding disk type is ext2. What do I have to do to allow me to backup larger chunks to tape? My tape drive is the Exabyte 8505 with a capacity of about 7 Gigabytes. When I list the /dosworld/pent14/e individual root directories in the disklist the data seems to get written to tape. My next step is to verify that the data on tape is in fact readable. Thanks, Vytas Vytas Janusauskas Dendron Resource Surveys Inc. 880 Lady Ellen Place, Suite 206 Ottawa, Ontario, Canada K1Z 5L9 Voice: (613) 725-2971 Fax: (613) 725-1716 mailto:[EMAIL PROTECTED]
