Hi--

This is probably easy for you old pros, but could not find enough in the man pages to figure it out.

1. How can I extract specific files and directories from a tarball? What is the exact command line command I need to enter, please?

2. Is there any way to speed up the process with a huge .tgz file?

3. If I want to restore all the files and all the rights and dates and times, do I need to be logged into X as root, or can I do it as a normal user?

My secondary HDD failed yesterday. My backups are stored in .tgz files. After replacing the drive, I used Red Hat 9.0's File Roller to get many of the files I need.

Those were from a tarball that was 1.5 gigs. The one I am having difficulty with is one that is 13 gigs. File Roller takes about 45 minutes to read the file so I can tell it what to grab. Then I started the extraction last night at 10:20 pm and sat up all night with it (I had other files I needed to extract as soon as that was done, and a deadline to meet); it still was not done at 5:20 am this morning when I stopped it to get at least enough of the other files to meet my deadline.

In any event, I want to pull out a couple of directories, let's call them /doug2 and /doug3, and a couple of files, /foobar/sub1/file1, and /foobar/sub2/file2.

I tried this (if I remember straight after this all-nighter! ):
tar -xvf /backups/weekly.tgz doug2
and it reported that there was no such directory found in that .tgz file.
I tried /doug2 and same result. Yet File Roller displayed /doug2. So I must be doing something wrong. What's the key wording here, please?

So, to speed the process, would it make sense for me to unzip the file first and then only untar the directory I want? Will it even be visible once it is unzipped? Do something else? This file only took about 45 minutes to create; why should it take 7 1/2 hours + to extract?

Something is up with your hardware that makes writes slower than reads, I guess.

tar -xvzf filename

Forgot to add:

you can use the "t" option to tar to list tarballs. eg tar tzvf foo.tar.gz

ugenn--

Thanks for the quick reply!

I checked and I actually used tar -xvzf.

How is that done? Not sure the specific language to use....

You suggested--
I had used--
So I need to remove the leading / in front of backups and I need to add /* after doug2, right? (normally I would just try these things, but it took it probably a half hour before it crapped out.

The hardware issue: I have a Celeron 2 ghz on an eMachines about 18 months old; the new drive is a Western Digital 120 gig, running as a slave with the jumper set to CS.

Thanks, ugenn!

You don't add the trailing /* because the shell will expand it. just doug2 will do. Since doug2 is a dir, tar will pull all the subdirs under it out as well.

Thanks ugenn!

I looked again and I think I reversed the order of the archive filename and the directory I wanted in the command line.

That is great help! Thanks!