no matches in file

[Julian Pace Ross]

Hi Wayne... the files I was dealing with before were files gzipped with the
--rsyncable argument. The original files were quite large, around 6GB, but I
decided to give it a go with these to keep things simple.
These are exports of a database (mssql) that are one day apart, and are
nearly identical in size, so one would imagine that the contents are fairly
similar.... right?

so anyway, I do a --no-whole-file local test:

 rsync -avvv --progress --no-whole-file --itemize-changes --stats
./pcutbak_prisma24/Data/Backup/pcutbak.bak
./pcutbak_prisma26/Data/Backup/pcutbak.bak

I get the following results: (would love a bit of help interpreting whether
the files are actually really that different, or maybe I can improve things
by changing the block size?)

total: matches=577  hash_hits=-502814583  false_alarms=133754
data=6871521032

and at the end..

Number of files: 1
Number of files transferred: 1
Total file size: 6919522816 bytes
Total transferred file size: 6919522816 bytes
Literal data: 6871521032 bytes
Matched data: 48001784 bytes
File list size: 29
File list generation time: 0.001 seconds
File list transfer time: 0.000 seconds
Total bytes sent: 6872362577
Total bytes received: 665655


Cheers
Julian




2009/2/4 Julian Pace Ross <linux at prisma.com.mt>

> It was a local copy, but I did indeed use --no-whole-file!! Any clue?
> the output shows below shows it: (delta-transmission enabled would be it i
> guess)
>
> 1 file to consider
> server_recv(2) starting pid=1820
> send_file_list done
> send_files starting
> recv_file_name(24\Data\Backup\pcutbak.bak)
> received 1 names
> recv_file_list done
> get_local_name count=1 /cygdrive/c/26\Data\Backup\pcutbak.bak
> generator starting pid=1820 count=1
> recv_files(1) starting
> delta-transmission enabled
> recv_generator(26\Data\Backup\pcutbak.bak,0)
> generating and sending sums for 0
> send_files(0, /cygdrive/c/24\Data\Backup\pcutbak.bak)
> count=32465 rem=22404 blength=32456 s2length=3 flength=1053673988
> send_files mapped /cygdrive/c/24\Data\Backup\pcutbak.bak of size 1053238900
> calling match_sums /cygdrive/c/24\Data\Backup\pcutbak.bak
> >f.st.... 24\Data\Backup\pcutbak.bak
> built hash table
> hash search b=32456 len=1053238900
> done hash search  99%   66.77Mbits/s    0:00:00
> sending file_sum
> false_alarms=11057 hash_hits=411107964 matches=0
>     1004.45MB  - 100%   59.89Mbits/s    0:02:14 (xfer#1, to-check=0/1)
> sender finished /cygdrive/c/24\Data\Backup\pcutbak.bak
> generate_files phase=1
> send_files phase=1
> recv_files(26\Data\Backup\pcutbak.bak)
>
> 2009/2/4 Wayne Davison <wayned at samba.org>
>
> On Tue, Feb 03, 2009 at 02:09:38PM +0100, Julian Pace Ross wrote:
>> > Hi, I have two versions of a file that in theory should be quite
>> similar...
>> > however, the transfer is taking more or less the same time as though the
>> > whole file is being transferred.
>>
>> I'd imagine that you did a local copy without using --no-whole-file
>> (since --whole-file is the default for more efficient local copying).
>>
>> ..wayne..
>>
>
>
-------------- next part --------------
HTML attachment scrubbed and removed