You want a holding disk.  Even though your tape is 'disk', taper can only
write to one disk file and without a holding disk, a single dumper has to
pipe to the taper.  You're limiting your throughput.  Taper is going to be
really fast so you only need about about 150% of your largest DLE for hold
space.  But you do need some.


Dana Bourgeois


> -----Original Message-----
> From: owner-amanda-users AT amanda DOT org 
> [mailto:owner-amanda-users AT amanda DOT org] On Behalf Of Turgut Kalfaoglu
> Sent: Monday, November 03, 2003 10:07 PM
> To: amanda-users AT amanda DOT org
> Subject: slow amanda performance on ONE system.
> 
> 
> 
> Hello - I am making great progress with Amanda; now backing 
> up three systems. (For some reason, samba backups did not 
> work for me; but installing the amanda client on our windows 
> server did the trick; I can backup over that). We have a 
> dedicated machine to do the backups, and it backs up two unix 
> servers (one Linux one SunOS 5.8), and a windows machine.
> 
> One puzzling thing is that our SunOS system seems to
> take a very long time making backups. I started a full backup 
> about 12 hours ago, and it's still running. If I wait long 
> enough, it finishes (I did it once before), but I would like 
> it to finish in a regular timeframe. There is no bottleneck 
> as far as I can tell; the machines are mostly idle; the 
> network connection is 100MB like the other machines. I am not 
> very good at reading the 'amstatus' output so I thought I 
> would ask for help from this very helpful group..
> 
> [amanda@yedek amanda]$ amstatus home | more
> Using /usr/adm/amanda/DailySet1/amdump from Mon Nov  3 
> 23:25:26 EET 2003
> 
> home:/etc            0     6296k finished (23:46:28)
> home:/usr/local      0   552408k finished (3:32:29)
> home:/usr/users      0  1516665k dumping to tape (3:32:29)
> home:/var/spool/mail 0  5028910k wait for dumping
> 
> SUMMARY          part      real  estimated
>                            size       size
> partition       :   4
> estimated       :   4              7227450k
> flush           :   0         0k
> failed          :   0                    0k           (  0.00%)
> wait for dumping:   1              5028910k           ( 69.58%)
> dumping to tape :   1              1516665k           ( 20.98%)
> dumping         :   0         0k         0k (  0.00%) (  0.00%)
> dumped          :   3   2075369k   2198540k ( 94.40%) ( 28.72%)
> wait for writing:   0         0k         0k (  0.00%) (  0.00%)
> wait to flush   :   0         0k         0k (100.00%) (  0.00%)
> writing to tape :   0         0k         0k (  0.00%) (  0.00%)
> failed to tape  :   0         0k         0k (  0.00%) (  0.00%)
> taped           :   2    558704k    681875k ( 81.94%) (  7.73%)
> 3 dumpers idle  : not-idle
> taper writing, tapeq: 0
> network free kps:      6570
> holding space   :         0k (  0.00%)
>  dumper0 busy   :  4:03:57  (100.00%)
>    taper busy   :  4:03:57  (100.00%)
>  0 dumpers busy :  0:00:00  (  0.00%)
>  1 dumper busy  :  4:03:57  (100.00%)            not-idle:  
> 4:03:57 (100.00%)
> 
> I just checked the 'home' system; the one that's being backed 
> up, and it shows that 'tar' is running, likewise for two 
> 'sendbackup' processes. 
> The last entry in /tmp/amanda belongs to amandad, and it has 
> not been updated for about 4 hours. it reads: CONNECT DATA 
> 921 MESG 922 INDEX 923 OPTIONS features=fffffeff9ffe0f;
> ----
> 
> amandad: time 0.133: got packet:
> ----
> Amanda 2.4 ACK HANDLE 000-A8E80608 SEQ 1067894728
> ----
> 
> amandad: time 0.133: pid 8343 finish time Tue Nov  4 03:29:37 2003
> 
> I guess this means that it finished.  
> 
> PS: I dont have a tape drive define; I backup to disk using 
> the very useful chg-disk "changer". Therefore I did not 
> define any holding disks.
> 
> I'd appreciate any feedback.. 
> Thanks, -turgut
> 
> 
> -----
> Turgut Kalfaoglu:  http://www.kalfaoglu.com
> EgeNet Internet Services: http://www.egenet.com.tr
> 
> 
>