Hi,
I ran some btape tests today to verify that I'd be improving throughput
by changing blocksize from 256KB to 2MB and found that this does indeed
appear to be true in terms of increasing compression efficiency, but it
doesn't seem to affect incompressible data much, if at all. Still, it
seems worth changing and I thank you for pointing me in that direction.
More importantly, I realized that my testing 6 months ago was not on all
4 of my drives, but only 2 of them. Today, I discovered one of my
drives (untested in the past) is getting 1/2 the throughput for random
data writes as the others!!
btape
*speed file_size=4 nb_file=4 skip_raw
SL500 Drive 0 SL500 Drive 1 C4 Drive 0 C4 Drive 1
256KB block size:
Zeros = 92.86 MB/s 92.36 MB/s 91.38 MB/s 92.86 MB/s
Random= 63.16 MB/s 27.53 MB/s 63.39 MB/s 63.60 MB/s
2MB block size:
Zeros = 123.5 MB/s 122.7 MB/s 122.7 MB/s 122.7 MB/s
Random= 62.24 MB/s 28.44 MB/s 63.62 MB/s 63.62 MB/s
----------------------------^
thanks,
Stephen
On 09/28/2012 05:08 AM, Alan Brown wrote:
> On 28/09/12 02:38, Stephen Thompson wrote:
>
>> Aren't these considered reasonable settings for LTO3?
>>
>> Maximum block size = 262144 # 256kb
>> Maximum File Size = 2gb
>
> Not really.
>
> Change maximum file size to 10Gb and maximum block size to 2M
>
> You _must_ set all open tapes to used and restart the storage daemon
> when changing the block size. Bacula can't cope with varying maximum
> sizes on a tape
>
> Even with those changes, if you have a lot of small, incompressible
> files you'll see high tape overheads.
>
>>
>>
>>
>> thanks for the help!
>> Stephen
>>
>
>
--
Stephen Thompson Berkeley Seismological Laboratory
stephen AT seismo.berkeley DOT edu 215 McCone Hall # 4760
404.538.7077 (phone) University of California, Berkeley
510.643.5811 (fax) Berkeley, CA 94720-4760
------------------------------------------------------------------------------
Got visibility?
Most devs has no idea what their production app looks like.
Find out how fast your code is with AppDynamics Lite.
http://ad.doubleclick.net/clk;262219671;13503038;y?
http://info.appdynamics.com/FreeJavaPerformanceDownload.html
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|