Topic: svn import for big file size showing error

[hasan]

Hello All,

I am trying to add a compressed file into a svn repository name "golden" on my sme-server.

After a long wait while it shows

Code: [Select]

Transmitting data...
It throws this error:

Code: [Select]

svn: PUT of '/golden/!svn/wrk/1460507f-ce71-4b38-8136-ef76289db0e8/ABC/datasets.tar.gz': Could not send request body: connection was closed by server (https://www.myserver.com) [/b]

Can anyone tell me the reason?
Is it because of file size restriction as the size of my tar.gz file is about 7GB?
If yes how can I adjust that, to resolve the issue?

 

Thanks and Regards
Hasan

[Stefano]

Hi

I'm not able to solve your problem but IMO:
- 7 GB is a BIG file
- multipost is not a good thing

Ciao
Stefano

[byte]

Moving this topic to the SME Server 7.x forum, it is more appropriate there. Thanks!

[cactus]

Hello All,

I am trying to add a compressed file into a svn repository name "golden" on my sme-server.

Why not add it uncompressed? In that case it would be a commit of smaller files (or you could commit in phases), which perhaps might work.

I do not know if there is a upper limit to commit size, and what number it has, but I can imagine that 7GB is very big. On top of that it is hard to judge from the small snippet of error log you post what the root cause of the error might be.

[hasan]

Thanks alot Cactus,

          Actually the uncompressed file is about a 40GB    folder containing text files of huge sizes.

So I do not want to put 40GB when I can put its compressed 7GB.

Thanks and Regards
Hasan

[cactus]

Thanks alot Cactus,

          Actually the uncompressed file is about a 40GB    folder containing text files of huge sizes.

So I do not want to put 40GB when I can put its compressed 7GB.

Subversion is optimized for working with text files, subversion is meant for version management. It has a clever mechanism to store the differences for each file with the previous version, for text files at least, I do not think it does that for binaries. By only storing the differences the footprint of the space used is will soon be smaller than for the individual versions combined.

If your data changes over time, as I guess, why else do you want to store it in a subversion repository, only the differences are saved and I therefore suggest storing the files as plain text and not as a tarball, my bet is the individual text files will also be smaller than 7Gb, right?