On 2013-02-05 00:06, Jim Henderson wrote:
> On Mon, 04 Feb 2013 22:26:01 +0000, sir chris wrote:
>
>> It was 33.2 GiB in 45 minutes. The second time (without -c) took 6
>> seconds.
>
> Yeah, the second time it would only pull changed data, so it would run a
> lot faster.
So that second run is only looking at timestamps, it would not serve as
a verify run for backups.
ā
Cheers / Saludos,
Carlos E. R.
(from 12.1 x86_64 āAsparagusā at Telcontar)
Normal file transfers on the other hand might wait until an entire file
> is transferred before performing an integrity check, and that means that
> a problem is discovered only after an entire file transfer has been
> attempted, and then would have to be attempted again.
No, rsync retransmits sections of files.
ā
Cheers / Saludos,
Carlos E. R.
(from 12.1 x86_64 āAsparagusā at Telcontar)
On Mon, 04 Feb 2013 23:38:06 +0000, Carlos E. R. wrote:
> On 2013-02-05 00:06, Jim Henderson wrote:
>> On Mon, 04 Feb 2013 22:26:01 +0000, sir chris wrote:
>>
>>> It was 33.2 GiB in 45 minutes. The second time (without -c) took 6
>>> seconds.
>>
>> Yeah, the second time it would only pull changed data, so it would run
>> a lot faster.
>
> So that second run is only looking at timestamps, it would not serve as
> a verify run for backups.
Correct. Though even if it checked checksums, itād still be faster than
doing a full backup, but it wouldnāt be as fast as just looking at file
size/timestamp. (IIRC, it looks at both when not using checksums)
On 2013-02-05 17:39, Jim Henderson wrote:
> On Mon, 04 Feb 2013 23:38:06 +0000, Carlos E. R. wrote:
>> So that second run is only looking at timestamps, it would not serve as
>> a verify run for backups.
>
> Correct. Though even if it checked checksums, itād still be faster than
> doing a full backup, but it wouldnāt be as fast as just looking at file
> size/timestamp. (IIRC, it looks at both when not using checksums)
Absolutely.
A verify run should mean a read of both source and destination files,
and create a checksum for both. I wonder if the filesystem metadata
could store the checksum, that would speed things up a lot. Or not⦠a
real checksum calculation is more reliable, of course.
ā
Cheers / Saludos,
Carlos E. R.
(from 12.1 x86_64 āAsparagusā at Telcontar)
Because the speed has a lot to do with the file size: it were 34.278 files, 2.026 sub-forlders.
I did the same (1st) command with the -c (rsync -r -t -v --progress -c -s) but all the same data is already copied.
Now the time was almost 24 minutes.
I also could send send a hwinfo file for more data over my system because that also has something to do with the speed (quite old system). But I donāt know if I then send the holy grail to break into my system, because it is a lot of data and I donāt want to read is all. :shame:
PS I find this program is more (or only) for backup because for normal copying its not very handy (teracopy is, as i said before it takes over the normal copy command in windows, teracopy canāt do quick check on size/timestamp).
There are more I would like to discuss should I put them here of search for similar treads or make a new one (slow linux vs win7: maybe a driver/hardware problem and I think internet browsing is slower and the system clock is changing the timeā¦).
On 02/07/2013 01:16 PM, sir chris wrote:
> There are more I would like to discuss should I put them here of search
> for similar treads or make a new one (slow linux vs win7: maybe a
> driver/hardware problem and I think internet browsing is slower and the
> system clock is changing the timeā¦).
each problem you want help with should be in its own thread, with a
descriptive subject (so the possible helpers can look in) and placed in
the correct sub-forum: