[quick] trying to split a large file but linux says it's to large
 

hi i'm trying to move servers and move all my data to the other server to i archived it up and it's a 90GB file but when i try to split it with a program like lxsplit i get the below error:

just tried the split command and it gave this error:

I assume that's a typo on your part (split: ytes=2048m)

What happens if you try

split -b 2097152 infile

it wasn't a typo that's just how the error was coming out with the "split –bytes=2048m file" command

that command you suggested is working but i added 3 more "0" so i get 2000mb files and not 2mb files but i have a question

i'm getting files like xaa, xab, xab etc and there will be about 50 of these but how do i join them back together?

Should have been

split –-bytes=2048m file (2 dashes)


use cat to rejoin

thanks for your help (again!) but there will be 50 randomly named files how do i join them with "cat" ?

There's nothing random in the files, you can set the prefix and the suffix as per man split.

cat xaa xab xac > outfile

will work, but you'll probably want to script a loop for 50 files.

hate to be a pain but how do i do that?

If all the chunks' filenames start with the lowercase letter x and are exactly three characters long then an example of how to do it might be:Of course, you'll probably have nothing but the chunks in the directory on the new box, and in such a case you wouldn't need to be specific at all:

Thanks - my looping is crap or I would have done it in my previous post.

thank you both very much!! when and split is done and i've moved them to my other server i'll try that loop command :)

No problem. BTW, make sure you checksum your file. :)

good idea.. how do i do that on linux? cksum filename right?

I once backed up my /home directory before installing SuSE 10.3. The resultant file would too large (30-40GB) for the vfat external drive I was backing up to. So I piped the output of tar through split and saved the slices on the external drive.

To restore, I mounted the drive, used cat and piped the output to the tar command. I didn't have to join the file first. I also used par2 to create parity files to enable the recovery of a damaged slice.

The examples near the end of the tar command has an example of moving a large directory using the -C option or a subshell on the right hand side of the pipe.

I'm not sure anout all the option here. I use md5sum, but there's also sha1sum and I would guess more.

Just have a look at man md5sum for example.

Do this before the split and then after the rejoining on the other machine

i'm trying to run "md5sum -c filename" but i'm getting rubbish.. is a 95gb file to big?

Have you already calculated the md5sum? The -c option is for checking.

Try

md5sum filename > filename.md5


To check you would do

md5sum -c filename.md5