Hi,
thank you very much for your help.
The real reason for the slowness was the networking card in the backup
server system.
It only was so strange that I did some scp copying and everything seemed to
work.
I did some backups that were perfectly okay. The backupduration were 2
minutes but just afterwards
the same backup lasted 40 minutes.
At the end I've seen a scp copy job presenting me "stalled". Then I started
to change the networking card
and now it's speeding.
So again thank you
Rgds
Ralf
Paul Bijnens
<paul.bijnens@xplan An: ralf.lebeda AT ezi DOT
de
ation.com> Kopie: amanda-users AT amanda
DOT org
Gesendet von: Thema: Re: Very, very slow
backup
owner-amanda-users@
amanda.org
26.01.2005 11:43
ralf.lebeda AT ezi DOT de wrote:
>
> what could be the reason for a very very slow backup?
Actually, I have no idea. Just a few remarks.
>
> My backup server is a celeron 500 with 128 MB memory and a tandberg slr
24
> tape box.
> My network is 100MB switched.
> The gamma client is a 800 MHz PIII system.
> The dv-02 client is a 1,2 GHz Athlon system.
My backup server is a celeron 300 MHz with 128 MB memory.
And I get a much larger throughput.
So it's not the servers capacity, neither the clients.
>
> I have no explanation for this slowness. If I copy a big file with scp
from
> gamma to backup server it is normal speed.
Also when you're copy from two different hosts at the same time?
A frequent error is a duplex mismatch on the switches; but that
would also be noticed when copying a large file from one host.
Just a tought...
Does "netstat -ni" have a large value in the error columns
on client and/or server?
> Bythe way the backupjob on gamm was only one iso image file.
Was that gnutar or dump?
Is the first time you make a backup? or was it fast before and
suddenly slow?
Do you have some clients where the problem does not exist?
Any strange message in the debug-files on client:
/tmp/amanda/*.<datetimestamp>.debug .
Was there any other activity on the clients?
Was there some non-responding nfs-directory on the clients (which
tar has to stat to find out it is an nfs-directory and skip it).
Is there anything else between the server and client? firewall
with misconfigured traffic-shaper, etc...?
When this backup is running, can you find out what the
dump/tar on the client is doing? (strace, truss, lsof...)
>
> Thanks and Rgds
> Ralf
>
>
>
> STATISTICS:
> Total Full Daily
> -------- -------- --------
> Estimate Time (hrs:min) 0:00
> Run Time (hrs:min) 26:59
> Dump Time (hrs:min) 51:53 51:53 0:00
> Output Size (meg) 1444.5 1444.5 0.0
> Original Size (meg) 1444.5 1444.5 0.0
> Avg Compressed Size (%) -- -- --
> Filesystems Dumped 2 2 0
> Avg Dump Rate (k/s) 7.9 7.9 --
>
> Tape Time (hrs:min) 0:16 0:16 0:00
> Tape Size (meg) 1444.5 1444.5 0.0
> Tape Used (%) 12.0 12.0 0.0
> Filesystems Taped 2 2 0
> Avg Tp Write Rate (k/s) 1517.3 1517.3 --
>
> USAGE BY TAPE:
> Label Time Size % Nb
> LXBACK127 0:16 1444.5 12.0 2
>
>
> NOTES:
> taper: tape LXBACK127 kb 1479232 fm 2 [OK]
>
>
> DUMP SUMMARY:
> DUMPER STATS TAPER STATS
> HOSTNAME DISK L ORIG-KB OUT-KB COMP% MMM:SS KB/s MMM:SS KB/s
> -------------------------- --------------------------------- ------------
> dv-02 -OpenOffice 0 821180 821180 -- 1502:31 9.1 9:011517.2
> gamma -amandatest 0 657970 657970 -- 1610:41 6.8 7:141517.4
Use the "columnspec" parameter in amanda.conf to add some spaces between
the columns.
>
> (brought to you by Amanda version 2.4.4p3)
>
>
>
--
Paul Bijnens, Xplanation Tel +32 16 397.511
Technologielaan 21 bus 2, B-3001 Leuven, BELGIUM Fax +32 16 397.512
http://www.xplanation.com/ email: Paul.Bijnens AT xplanation DOT com
***********************************************************************
* 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 *
***********************************************************************
|
