0

We have ~10K tar files which are using ~30TB of disk space. For all tar files, BLOBs are stored inside 2 subdirectories. First directory is the user then date of the blob created then bunch of blobs.

Each blob is a directory containing single directory with some files. We are using windows 8 server, I need to extract all these tar files and blobs to restore the data.

I have to untar first then deblob the files. I would like to avoid the number of read/write file operations and utilize the space efficiently.

Is there a way I can open tar file in memory and deblob the file and write the final directory structure?

enter image description here

Anil Kumar C
  • 1,604
  • 4
  • 22
  • 43
  • People suggest using Ant libraries, where tar is implemented http://stackoverflow.com/questions/315618/how-do-i-extract-a-tar-file-in-java – Alex Nevidomsky Apr 21 '15 at 16:47
  • Two levels of uncompressing will not help us. if we do that, we have to accommodate ~120TB. We have only 60TB available. We need to read a tar file extracted it on the memory and write all the files in one operation. – Anil Kumar C Apr 21 '15 at 17:04
  • 1
    What is the "deblob" operation? Because you can get an `InputStream` for every file in the tar with suitable libraries (e.g. jtar) without having the extract to the tar archive's contents to disk. – dhke Apr 21 '15 at 17:25
  • @dhke- Not sure how it works, I will try with a PoC. Thanks for the suggestion. – Anil Kumar C Apr 22 '15 at 13:51

0 Answers0