is current armv7a_hardfp stage tarball corrupted?

[duby2291]

I'm trying to untar the current stage tarball to an android device but tar is crapping out with error messages. I've tryed using both gnu tar and busybox tar, both give similar messages and stop at exactly the same spot. When it gets to extracting /usr/libexec/gcc/armv7a-hardfloat-linux-gnueabi/4.7.3/f951 file is where it craps out. Using busbox tar the error is "tar: bunzip error -5 tar: short read". Using gnu tar the error is "tar: Unexpected EOF in archive tar: Error is not recoverable: exiting now"

If anyone has any suggestions, i'd really like to hear some opinions on what is going on. Thanks everyone.

EDIT: Exactly the same thing happens when I try the latest funtoo stage tarball for armv7a_hardfp. Exactly the same error on exactly the same file during extraction

[cwr]

Is it perhaps a space problem? The 20130209 version is ok.

Will

[duby2291]

No, I have a 16gig sdcard with an ext2 partition, so space shouldn't be the problem at all. Honestly I can't think of a single reason why it won't extract. It seems perfectly legitimate and it should wiork just fine, but something is going on that I don't know about that is preventing the file from extracting. I'm really not sure what is going on.

[duby2291]

Ok, so I got access to a pc and i plugged my sdcard into a USB adapter and it extracted fine from there., later on as i was extracting the portage snapshot I ran out of inodes. So I reformatted the partition with a smaller inode size. I no longer have access to that pc and am attempting to extract the stage tarball again, buf I am still getting the same exact problem. Something about android is messing this up, but i don't know where to loo,. Anybody have any idea's?

I know that it extracts fine from a pc, but it won't extract from androiddue to the problem I described in the first post.

[cwr]

Well, I've run into inode problems with portage on small devices (< 2G) before now.
"df -i" is happy, right?

Will

[duby2291]

Ok, so the partition looks good now. I had to reformat it, but it looks like everything is in order now.

[cwr]

Looking at my /usr/portage, which is on its own partition, I get:

Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/sda10 1050576 165833 884743 16% /usr/portage

so allowing for 5.5K files in /usr/portage/distfiles you should be fine with
160K inodes; you appear to have more than enough.

Have you tried unpacking the file on another machine and mounting that
partition via NFS or CIFS?

Will

[duby2291]

I did manage to get the file unpacked from my pc running gentoo. I just have to plug the sdcard into a USB adapter and mount the partition. But the same file wont extract from the phone itself.

So i did get the file extracted, and i got the partition formated with the smallest inode size possible, and i managed to get the portage tree extracted. I even managed to chroot into it.....

But, now i'm running out of inodes during emerge -e world.....

This shit sucks. There has to be a better way to distribute the portage tree. It has way too many small files. I think i'm out of options, there doesn't seem to be any more options to make this work.

[duby2291]

oops double post...

[cwr]

I don't understand why the tar file is unpacking on one system and not another,
when there is no shortage of inodes. One last shot might be to format with
eg: mkfs /dev/xxx -t yyy -T news, which quadruples the number of inodes
(see /etc/mke2fs.conf).


Will