Re: Using only one tape, full dump delayed
Ok Guys You're so smart and kind. I do hope 'dude' is not too rude (since it's not my native language). I found it after disassembling the drive (some of you may have a notch at the bottom of the case, the SUN-Version does not). BTW it does work fine for days already, and I like it. Cheers Urs On Mon, Mar 03, 2003 at 04:39:05PM +0100, Urs Forster wrote: The '72-144 Gbyte 4mm DDS-3 Autoloader Installation and User's Guide' says that a DDS-3 125m tape would hold 72GB uncompressed and 144GB compressed. Dude, that's the capacity of the whole *Library*! 6*12GB = 72GB, 6*24GB = 144GB. The tape itselfes says: 12GB native Capacity. That is correct, for a a DDS-3 tape drive with 125m tape. There is no switch inside the drive to enable/disable compression, Really? Not like this? http://h2.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=lpg51068 so I assume when not using the 'c,u' option, it wont compress. You have some sort of front-panel display that will indicate compression during data write, don't you? Then there is the density switch (l,m,h) for low, medium and high. I don't think this has to do with compression. I believe this depends on the made of the tape? (I am not using any) HP has what seems to be a pretty good Solaris install doc here: http://h2.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=lpg51069 -- Jay Lessert [EMAIL PROTECTED] Accelerant Networks Inc. (voice)1.503.439.3461 Beaverton OR, USA(fax)1.503.466.9472
Re: Using only one tape, full dump delayed
On Mon, Mar 03, 2003 at 04:39:05PM +0100, Urs Forster wrote: Dear Jon Since you found out anyway, I give up lying; I gonna tell you frankly: I don't know exactly what I have. 'dmesg' says: st18: HP DDS-3 4mm DAT loader 'inquire' says: [EMAIL PROTECTED]:HP C1557A U709|Tape, /dev/rmt/0mbn OK, so the same unit I have. I did believe that this is similar to a Surestore 12000e. Is there a 12000? I think the 1200 is a DDS-2. Also I remember that somewhere it said the 'b' (berkley) option was a good idea for amanda. I trust my sources and experience to skip the 'b'. The '72-144 Gbyte 4mm DDS-3 Autoloader Installation and User's Guide' says that a DDS-3 125m tape would hold 72GB uncompressed and 144GB compressed. Not a tape, but the entire magazine of 6 tapes. The tape itselfes says: 12GB native Capacity. And 6x12 == 72. There is no switch inside the drive to enable/disable compression, You're Lying again :)) so I assume when not using the 'c,u' option, it wont compress. With my internal switch settings to power up in no compression, allow software selection, and the driver in my Solaris 89 system, c/u give HW compression, h seems to be the same, at least I have not detected any difference, l gives HW compression off, and m gives leave it as it was. Then there is the density switch (l,m,h) for low, medium and high. I don't think this has to do with compression. I believe this depends on the made of the tape? (I am not using any) Sorry about the 'n' option. That was a lie ;-) What do you mean by made of the tape? make? mode? ??? Get to the HP support pages for the 24x6 surestore changers. Look for manuals and the like. Specifically for anything to do with installation and unix. They have very distinct setting suggestions for several flavors of unix for both the back panel switches and that 8 place, non-existant, internal DIP switch. If you follow their settings AND edit /kernel/drv/st.conf to look at lun 1 as well as lun 0, then you will reconfigure reboot into two /dev/rmt devices, one for the tape drive, one for the changer. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: Using only one tape, full dump delayed
Hi ok, with that data given i would suggest the following: adjust your tapetype to match the size of your tapes, set runtapes to 1, leave tapecycle at 6, leave dumpcycle at one week and runspercycle at 5, and amanda will do nicely what you want her to do. Your only problem is the wrong length of your tapes in tapetype. Christoph [EMAIL PROTECTED] schrieb: Hi, now let's get some information to sort out varios things: 1.) what length are the tapes you use: 120 meter or 125 meter? 125m / 410 feet (12.0GB) 2.) how often do you want amanda to take level0 dumps of your data? once a week? Does not matter, as long as at least one full dump is on one of the 6 tapes... 3.) how many data do you need to backup? ~15GB presently (2 Machines, 12 Partitions). Will not grow much. 4.) is hardwarecompression active in your tapedrive? No, disabled by using device /dev/rmt/0bn (b= berkley mode n= no compression) Urs Christoph
Re: Using only one tape, full dump delayed
Hello Christoph Thanks a lot for advise. The only problem - as you call it - is easily solved ;-) Let's see what she says Cheers Urs Hi ok, with that data given i would suggest the following: adjust your tapetype to match the size of your tapes, set runtapes to 1, leave tapecycle at 6, leave dumpcycle at one week and runspercycle at 5, and amanda will do nicely what you want her to do. Your only problem is the wrong length of your tapes in tapetype. Christoph [EMAIL PROTECTED] schrieb: Hi, now let's get some information to sort out varios things: 1.) what length are the tapes you use: 120 meter or 125 meter? 125m / 410 feet (12.0GB) 2.) how often do you want amanda to take level0 dumps of your data? once a week? Does not matter, as long as at least one full dump is on one of the 6 tapes... 3.) how many data do you need to backup? ~15GB presently (2 Machines, 12 Partitions). Will not grow much. 4.) is hardwarecompression active in your tapedrive? No, disabled by using device /dev/rmt/0bn (b= berkley mode n= no compression) Urs Christoph
Re: Using only one tape, full dump delayed
Here is what I put to config.
I have no idea what that filemark is. So if its wrong, I am not lying ;)
define tapetype SURESTORE-1200E {
comment HP AutoLoader
length 12000 mbytes
filemark 100 kbytes
speed 500 kbytes
}
I believe the paramaters you give are not the real parameters
of the config that is running, or I must be very stupid.
Ok, you got me. I never said it's 3.9G. Actually the tape holds 12.0G.
Sorry Guys, I'm not joking, just trying hard.
The config below was just the example and I did not change it :-(
Did not change in the mail? Or did not change it the Config?
It's better that you don't lie to Amanda. It works much better if
she has more correct info.
--
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 *
***
Re: Using only one tape, full dump delayed
On Fri, Feb 28, 2003 at 11:06:56AM +0100, [EMAIL PROTECTED] wrote:
Here is what I put to config.
I have no idea what that filemark is. So if its wrong, I am not lying ;)
define tapetype SURESTORE-1200E {
comment HP AutoLoader
length 12000 mbytes
filemark 100 kbytes
speed 500 kbytes
}
Actually I suspect you have been lying in a number of places and ways.
Perhaps unintentionally :))
What tape drive do you really have?
If I recall correctly, the SureStore-1200 is a DDS2 (DAT-2) changer
and you were using a reasonable (3.9GB) tapetype definition for that.
But you were using 125M tapes, DDS3 tapes that should not work in a
DDS2 drive. Actually, has anyone ever tried that?
And for sure, even using a 5M longer DDS3 tape with a DDS2 drive, I
don't think would not get anywhere near the 8GB capacity you seem to
be getting.
Unless you can write to DDS3 formula tapes in a DDS2 drive AND we
consider your other lie :)
4.) is hardwarecompression active in your tapedrive?
No, disabled by using device /dev/rmt/0bn
(b= berkley mode n= no compression)
Based on device name I'm guessing this is a Solaris or HP system.
In either case, n refers to No Rewind, not no compression.
By chosing the 0 device with no density modifier (like l, m, ...),
you have elected default compression, which generally is on.
So maybe, if you are using a DDS2 drive, a DDS3 tape, no software
compression, and hardware compression, you can get 8+GB of data written.
What are you actually doing?
BTW using the Berkeley drive options is not recommended.
--
Jon H. LaBadie [EMAIL PROTECTED]
JG Computing
4455 Province Line Road(609) 252-0159
Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: Using only one tape, full dump delayed
On Thu, Feb 27, 2003 at 10:20:41AM +0100, [EMAIL PROTECTED] enlightened us:
Hi all
amanda keeps using only one tape per day but always delays full dump for
the same disk
(/dev/dsk/c0t0d0s08758612 6866287 180473980%/ora)
Why cant it use a second tape?
There are 6 Tapes in the changer;
dumpcycle 1 weeks
runspercycle 5
tapecycle 6 tapes
runtapes 3
tapetype SURESTORE-1200E
define tapetype SURESTORE-1200E {
comment HP AutoLoader
length 3900 mbytes
filemark 100 kbytes
speed 500 kbytes
}
What's wrong whith it?
snip
NOTES:
planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger than tape,
full dump delayed]
This is why. It must be bigger than 3900 mbytes after amanda compresses it,
so it won't fit on a single tape. Amanda doesn't span tapes, so you'll have
to split this up with tar.
--
Matt Hyclak
Department of Mathematics
Ohio University
(740) 593-1263
pgp0.pgp
Description: PGP signature
Re: Using only one tape, full dump delayed
From: Christoph Scheeder [EMAIL PROTECTED] Hi, first : simple answer, simple question, and amanda is telling it to you: amanda thinks the level 0 dump of this filesystem will not fit on a single tape. That's understood. Amanda can not span a single disk-list entry over multiple tapes. you'll have to split it up using excludelists and tar. [...] NOTES: planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger than tape, full dump delayed] [...] second: these values will give you problems. There are 6 Tapes in the changer; dumpcycle 1 weeks runspercycle 5 tapecycle 6 tapes runtapes 3 you want 5 runs per week, each using up to 3 tapes. This tells amanda: do a level 0 for each filesystem at least every 15 tapes. but a little later you tell her she only has 6 tapes in rotation. this is a NoNo with amanda. either lower your dump-cycle or get more tapes. runtapes should be at least one tape bigger then runspercycle x runtapes, or amanda will get confused about when to do a level 0 of your disks. OK, I see the problem, but not the solution yet ;-) There are 6 Tapes in the changer and I dont want to swap these. It is supposed to be sort of unmaintained forever. - If I put runtapes 1: will it still use more than one tape if needed? - If I put runspercycle 3 and runtapes 2 then I still would need 7 Tapes... - If I put runspercycle 2 and still call it 5 times: will it confuse amanda? I know I could find out trying but it would take ages. if you want only level 0 backups, you should set dumpcycle to 0. Christoph
Re: Using only one tape, full dump delayed
I believe the paramaters you give are not the real parameters
of the config that is running, or I must be very stupid.
[EMAIL PROTECTED] wrote:
...
tapetype SURESTORE-1200E
define tapetype SURESTORE-1200E {
comment HP AutoLoader
length 3900 mbytes
filemark 100 kbytes
speed 500 kbytes
}
So you have tapes that have 3900 Mbytes capacity...
What's wrong whith it?
Cheers
Urs
Here's the amanda report:
These dumps were to tape Daily-06.
Amanda needed only one tape.
The next 3 tapes Amanda expects to used are: Daily-01, Daily-02, Daily-03.
FAILURE AND STRANGE DUMP SUMMARY:
pbrazor1.t /var lev 0 STRANGE
pbrazor1.t /razor lev 0 STRANGE
STATISTICS:
Total Full Daily
Estimate Time (hrs:min)0:06
Run Time (hrs:min)12:44
Dump Time (hrs:min) 13:02 1:43 11:18
Output Size (meg)8414.6 1594.8 6819.8
How is it possible that Amanda cat put 8414 Mbyte (already compressed?)
on one tape of 3900 Mbyte?
Original Size (meg) 13959.4 3912.710046.7
Avg Compressed Size (%)60.3 40.8 67.9 (level:#disks ...)
Filesystems Dumped 12 10 2 (1:1 2:1)
Avg Dump Rate (k/s) 183.7 263.3 171.6
Tape Time (hrs:min) 11:51 0:33 11:18
Tape Size (meg) 8415.1 1595.2 6819.8
Again, amanda has put 8415 Mbyte on your tape that apparently holds only
3900 Mbyte???
Tape Used (%) 215.8 40.9 174.9 (level:#disks ...)
Filesystems Taped12 10 2 (1:1 2:1)
Avg Tp Write Rate (k/s) 202.0 834.0 171.6
FAILED AND STRANGE DUMP DETAILS:
...
NOTES:
planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger than tape,
full dump delayed]
Here is the reason: Amanda believes it cannot hold a full dump on one
tape. That's a limitation of Amanda: one DLE (disk list entry) has to
fit on one tape.
As an easy workaround, you can make DLE's of subdirectories instead, if
the directory contains data that can be backed up separatly, for example
the oracle software as one DLE and the datafiles directory as another
DLE (but for the datafiles, you better choose some other method, like
making a hot backup with your tablespaces in backup mode to another
directory and make an amanda backup thereof).
...
taper: tape Daily-06 kb 8617056 fm 12 [OK]
And indeed, it's very strange that amanda can put 8617056 Kbytes on a
tape that you say can only hold 3900*1024 Kbytes.
I suppose that the real config that your amanda runs has uses other values.
DUMP SUMMARY:
DUMPER STATSTAPER STATS
HOSTNAME DISKL ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s
-- -
mambo.tszrh. / 0 1475010 493440 33.5 22:28 366.1 9:42 847.5
mambo.tszrh. /ebs0 110670 15008 13.6 1:10 213.4 0:39 388.5
mambo.tszrh. /var0 253400 166336 65.6 4:43 588.2 3:10 873.6
pbrazor1.tsz / 0 124540 49792 40.0 4:02 205.7 0:59 849.3
pbrazor1.tsz -xport/home 0 375170 131872 35.1 10:27 210.4 2:43 811.5
pbrazor1.tsz /opt0 564690 195136 34.6 13:25 242.2 3:48 855.6
pbrazor1.tsz /ora2 68541004602336 67.1 571:52 134.1 571:53 134.1
Your 3900 Mbyte tape holds one image of 4600 Mbyte...
pbrazor1.tsz /razor 0 0 150976 -- 13:04 192.4 3:03 823.2
pbrazor1.tsz /razorshr 1 34337302381184 69.3 106:21 373.2 106:22 373.1
pbrazor1.tsz /usr0 491760 192352 39.1 14:11 225.9 3:45 854.9
pbrazor1.tsz /usr/local 0 611370 195776 32.0 15:50 206.0 3:56 830.0
pbrazor1.tsz /var0 0 42848 --4:03 176.5 0:54 795.1
(brought to you by Amanda version 2.4.3)
But if I'm proved wrong, the whole backup community would be very
interested in how you can put that much data on such small tapes...
You could make a fortune here, if you start you own company!
--
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 *
***
Re: Using only one tape, full dump delayed
On Thu February 27 2003 04:20, [EMAIL PROTECTED] wrote:
Hi all
amanda keeps using only one tape per day but always delays full
dump for the same disk
(/dev/dsk/c0t0d0s08758612 6866287 180473980%/ora)
If the above is in bytes, I'm stumped, if its in kilobytes, it looks
as if its too big for one tape at 8,758.612 megs.
Since amanda cannot span a single DLE across more than 1 tape, she
is apparently refusing to do it.
You will need to break that DLE down into subdir listings, such as,
/dev/dsk/c0t0d0s0/home
/dev/dsk/c0t0d0s0/etc
/dev/dsk/c0t0d0s0/whatever
in order to get its present estimated size of 8,758,612 broken down
into pieces that will fit on a tape. Then when it hits EOT while
writing one of the smaller DLE's, it should unload that tape, load
the next and restart that unfinished DLE from its beginning on the
next tape.
If it refuses to do the smaller entries, then there must be a
permissions problem of some kind. From the other warnings issued
it looks as if you are doing samba shares instead of installing the
amanda clients on those machines. I tried that for about a month,
and gave up on using samba which has gotchas going clear back to
dos days. I was backing up everything every time, wasting lots of
tape space because it doesn't support the file time data in a
compatible way.
Why cant it use a second tape?
There are 6 Tapes in the changer;
dumpcycle 1 weeks
runspercycle 5
tapecycle 6 tapes
runtapes 3
tapetype SURESTORE-1200E
define tapetype SURESTORE-1200E {
comment HP AutoLoader
length 3900 mbytes
filemark 100 kbytes
speed 500 kbytes
}
What's wrong whith it?
Cheers
Urs
Here's the amanda report:
These dumps were to tape Daily-06.
The next 3 tapes Amanda expects to used are: Daily-01, Daily-02,
Daily-03.
FAILURE AND STRANGE DUMP SUMMARY:
pbrazor1.t /var lev 0 STRANGE
pbrazor1.t /razor lev 0 STRANGE
STATISTICS:
Total Full Daily
Estimate Time (hrs:min)0:06
Run Time (hrs:min)12:44
Dump Time (hrs:min) 13:02 1:43 11:18
Output Size (meg)8414.6 1594.8 6819.8
Original Size (meg) 13959.4 3912.710046.7
Avg Compressed Size (%)60.3 40.8 67.9
(level:#disks ...) Filesystems Dumped 12 10
2 (1:1 2:1) Avg Dump Rate (k/s) 183.7 263.3
171.6
Tape Time (hrs:min) 11:51 0:33 11:18
Tape Size (meg) 8415.1 1595.2 6819.8
Tape Used (%) 215.8 40.9 174.9
(level:#disks ...) Filesystems Taped12 10
2 (1:1 2:1) Avg Tp Write Rate (k/s) 202.0 834.0
171.6
FAILED AND STRANGE DUMP DETAILS:
/-- pbrazor1.t /var lev 0 STRANGE
sendbackup: start [pbrazor1.tszrh.csfb.com:/var level 0]
sendbackup: info BACKUP=/usr/local/bin/gtar
sendbackup: info RECOVER_CMD=/usr/local/bin/gzip -dc
|/usr/local/bin/gtar -f... -
sendbackup: info COMPRESS_SUFFIX=.gz
sendbackup: info end
? gtar: ./cron/olog: file changed as we read it
? gtar: ./tmp/wsconAAAkITF9_:0.0: file changed as we read it
| Total bytes written: 193341440
? gtar: Error exit delayed from previous errors
??error [/usr/local/bin/gtar returned 2]? dumper: strange [missing
size line from sendbackup]
? dumper: strange [missing end line from sendbackup]
\
/-- pbrazor1.t /razor lev 0 STRANGE
sendbackup: start [pbrazor1.tszrh.csfb.com:/razor level 0]
sendbackup: info BACKUP=/usr/local/bin/gtar
sendbackup: info RECOVER_CMD=/usr/local/bin/gzip -dc
|/usr/local/bin/gtar -f... -
sendbackup: info COMPRESS_SUFFIX=.gz
sendbackup: info end
? gtar: ./apache/logs/access_log: file changed as we read it
| Total bytes written: 937625600
? gtar: Error exit delayed from previous errors
??error [/usr/local/bin/gtar returned 2]? dumper: strange [missing
size line from sendbackup]
? dumper: strange [missing end line from sendbackup]
\
NOTES:
planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger
than tape, full dump delayed]
planner: Full dump of pbrazor1.tszrh.csfb.com:/export/home
promoted from 2 days ahead.
planner: Full dump of pbrazor1.tszrh.csfb.com:/razor promoted
from 2 days ahead.
planner: Full dump of pbrazor1.tszrh.csfb.com:/usr/local
promoted from 2 days ahead.
planner: Full dump of mambo.tszrh.csfb.com:/var promoted from 2
days ahead. planner: Full dump of mambo.tszrh.csfb.com:/ebs
promoted from 2 days ahead. planner: Full dump of
pbrazor1.tszrh.csfb.com:/var promoted from 2 days ahead.
planner: Full dump of pbrazor1.tszrh.csfb.com:/opt promoted from
2 days ahead.
planner: Full dump of pbrazor1.tszrh.csfb.com:/usr promoted from
2 days ahead.
planner: Full dump of mambo.tszrh.csfb.com:/ promoted from 2
days ahead. planner: Full dump of pbrazor1.tszrh.csfb.com:/
promoted from 2 days ahead. taper: tape Daily-06 kb 8617056 fm 12
[OK]
DUMP SUMMARY:
Re: Using only one tape, full dump delayed
On Thu February 27 2003 07:53, Matt Hyclak wrote:
On Thu, Feb 27, 2003 at 10:20:41AM +0100, [EMAIL PROTECTED]
enlightened us:
Hi all
amanda keeps using only one tape per day but always delays full
dump for the same disk
(/dev/dsk/c0t0d0s08758612 6866287 180473980%/ora)
Why cant it use a second tape?
There are 6 Tapes in the changer;
dumpcycle 1 weeks
runspercycle 5
tapecycle 6 tapes
runtapes 3
tapetype SURESTORE-1200E
define tapetype SURESTORE-1200E {
comment HP AutoLoader
length 3900 mbytes
filemark 100 kbytes
speed 500 kbytes
}
What's wrong whith it?
snip
NOTES:
planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger
than tape, full dump delayed]
This is why. It must be bigger than 3900 mbytes after amanda
compresses it, so it won't fit on a single tape. Amanda doesn't
span tapes, so you'll have to split this up with tar.
What am I missing here Matt? Thats not the same DLE he quotes at
the top of the message, so how does that become reconciled?
Not trying to be nibby, just curious :)
--
Cheers, Gene
AMD [EMAIL PROTECTED] 320M
[EMAIL PROTECTED] 512M
99.23% setiathome rank, not too shabby for a WV hillbilly
Re: Using only one tape, full dump delayed
On Thu, Feb 27, 2003 at 10:43:31AM -0500, Gene Heskett enlightened us:
On Thu February 27 2003 07:53, Matt Hyclak wrote:
On Thu, Feb 27, 2003 at 10:20:41AM +0100, [EMAIL PROTECTED]
enlightened us:
Hi all
amanda keeps using only one tape per day but always delays full
dump for the same disk
(/dev/dsk/c0t0d0s08758612 6866287 180473980%/ora)
Why cant it use a second tape?
There are 6 Tapes in the changer;
dumpcycle 1 weeks
runspercycle 5
tapecycle 6 tapes
runtapes 3
tapetype SURESTORE-1200E
define tapetype SURESTORE-1200E {
comment HP AutoLoader
length 3900 mbytes
filemark 100 kbytes
speed 500 kbytes
}
What's wrong whith it?
snip
NOTES:
planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger
than tape, full dump delayed]
This is why. It must be bigger than 3900 mbytes after amanda
compresses it, so it won't fit on a single tape. Amanda doesn't
span tapes, so you'll have to split this up with tar.
What am I missing here Matt? Thats not the same DLE he quotes at
the top of the message, so how does that become reconciled?
Not trying to be nibby, just curious :)
There's only one /ora in his dump summary...I was working on the assumption
that the output from df at the beginning of the message is the same
partition. It looks to me that it is an ~8.7GB partition with ~6.8GB in use
and ~1.8GB free. Obviously 6.8GB isn't going to fit on a 4GB tape, but
compressed it might have at one point in time. It's just not anymore.
So the way I see it he'll have to split /ora up with tar, no?
Matt
--
Matt Hyclak
Department of Mathematics
Ohio University
(740) 593-1263
pgp0.pgp
Description: PGP signature
Re: Using only one tape, full dump delayed
I believe the paramaters you give are not the real parameters
of the config that is running, or I must be very stupid.
Ok, you got me. I never said it's 3.9G. Actually the tape holds 12.0G.
Sorry Guys, I'm not joking, just trying hard.
The config below was just the example and I did not change it :-(
[EMAIL PROTECTED] wrote:
...
tapetype SURESTORE-1200E
define tapetype SURESTORE-1200E {
comment HP AutoLoader
length 3900 mbytes
filemark 100 kbytes
speed 500 kbytes
}
Re: Using only one tape, full dump delayed
Hi, now let's get some information to sort out varios things: 1.) what length are the tapes you use: 120 meter or 125 meter? 2.) how often do you want amanda to take level0 dumps of your data? once a week? 3.) how many data do you need to backup? 4.) is hardwarecompression active in your tapedrive? Christoph [EMAIL PROTECTED] schrieb: From: Christoph Scheeder [EMAIL PROTECTED] Hi, first : simple answer, simple question, and amanda is telling it to you: amanda thinks the level 0 dump of this filesystem will not fit on a single tape. That's understood. Amanda can not span a single disk-list entry over multiple tapes. you'll have to split it up using excludelists and tar. [...] NOTES: planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger than tape, full dump delayed] [...] second: these values will give you problems. There are 6 Tapes in the changer; dumpcycle 1 weeks runspercycle 5 tapecycle 6 tapes runtapes 3 you want 5 runs per week, each using up to 3 tapes. This tells amanda: do a level 0 for each filesystem at least every 15 tapes. but a little later you tell her she only has 6 tapes in rotation. this is a NoNo with amanda. either lower your dump-cycle or get more tapes. runtapes should be at least one tape bigger then runspercycle x runtapes, or amanda will get confused about when to do a level 0 of your disks. OK, I see the problem, but not the solution yet ;-) There are 6 Tapes in the changer and I dont want to swap these. It is supposed to be sort of unmaintained forever. - If I put runtapes 1: will it still use more than one tape if needed? - If I put runspercycle 3 and runtapes 2 then I still would need 7 Tapes... - If I put runspercycle 2 and still call it 5 times: will it confuse amanda? I know I could find out trying but it would take ages. if you want only level 0 backups, you should set dumpcycle to 0. Christoph
Re: Using only one tape, full dump delayed
Hi, now let's get some information to sort out varios things: 1.) what length are the tapes you use: 120 meter or 125 meter? 125m / 410 feet (12.0GB) 2.) how often do you want amanda to take level0 dumps of your data? once a week? Does not matter, as long as at least one full dump is on one of the 6 tapes... 3.) how many data do you need to backup? ~15GB presently (2 Machines, 12 Partitions). Will not grow much. 4.) is hardwarecompression active in your tapedrive? No, disabled by using device /dev/rmt/0bn (b= berkley mode n= no compression) Urs Christoph [EMAIL PROTECTED] schrieb: From: Christoph Scheeder [EMAIL PROTECTED] Hi, first : simple answer, simple question, and amanda is telling it to you: amanda thinks the level 0 dump of this filesystem will not fit on a single tape. That's understood. Amanda can not span a single disk-list entry over multiple tapes. you'll have to split it up using excludelists and tar. [...] NOTES: planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger than tape, full dump delayed] [...] second: these values will give you problems. There are 6 Tapes in the changer; dumpcycle 1 weeks runspercycle 5 tapecycle 6 tapes runtapes 3 you want 5 runs per week, each using up to 3 tapes. This tells amanda: do a level 0 for each filesystem at least every 15 tapes. but a little later you tell her she only has 6 tapes in rotation. this is a NoNo with amanda. either lower your dump-cycle or get more tapes. runtapes should be at least one tape bigger then runspercycle x runtapes, or amanda will get confused about when to do a level 0 of your disks. OK, I see the problem, but not the solution yet ;-) There are 6 Tapes in the changer and I dont want to swap these. It is supposed to be sort of unmaintained forever. - If I put runtapes 1: will it still use more than one tape if needed? - If I put runspercycle 3 and runtapes 2 then I still would need 7 Tapes... - If I put runspercycle 2 and still call it 5 times: will it confuse amanda? I know I could find out trying but it would take ages. if you want only level 0 backups, you should set dumpcycle to 0. Christoph
Re: Using only one tape, full dump delayed
On Thu February 27 2003 11:05, Matt Hyclak wrote: On Thu, Feb 27, 2003 at 10:43:31AM -0500, Gene Heskett enlightened us: On Thu February 27 2003 07:53, Matt Hyclak wrote: On Thu, Feb 27, 2003 at 10:20:41AM +0100, [EMAIL PROTECTED] [...] NOTES: planner: pbrazor1.tszrh.csfb.com /ora 20030226 0 [dump larger than tape, full dump delayed] This is why. It must be bigger than 3900 mbytes after amanda compresses it, so it won't fit on a single tape. Amanda doesn't span tapes, so you'll have to split this up with tar. What am I missing here Matt? Thats not the same DLE he quotes at the top of the message, so how does that become reconciled? Not trying to be nibby, just curious :) There's only one /ora in his dump summary...I was working on the assumption that the output from df at the beginning of the message is the same partition. It looks to me that it is an ~8.7GB partition with ~6.8GB in use and ~1.8GB free. Obviously 6.8GB isn't going to fit on a 4GB tape, but compressed it might have at one point in time. It's just not anymore. So the way I see it he'll have to split /ora up with tar, no? Yes. :-) -- Cheers, Gene AMD [EMAIL PROTECTED] 320M [EMAIL PROTECTED] 512M 99.24% setiathome rank, not too shabby for a WV hillbilly
