4

While researching this I have found a lot of messages about this concerning the exhaustion of memory, but I am not actually getting any malloc errors.

When I try and clone I get this mid-compression

error: git-upload-pack: git-pack-objects died with error.
fatal: git-upload-pack: aborting due to possible repository corruption on the remote side.
remote: aborting due to possible repository corruption on the remote side.
fatal: early EOF
fatal: index-pack failed

When I go to the remote and run git gc it gives me this.. (same amount through compression)

error: failed to run repack

When I run git fsck I get no output at all...

ideas?

jondavidjohn
  • 61,812
  • 21
  • 118
  • 158

1 Answers1

0

It depends on the line you should see before git-pack-objects died with error.

For instance, if you should see a line like:

remote: fatal: object 21f3981dd35fccd28febabd96f27241eea856c50 is corrupted

(as described in "GIT pull error - remote object is corrupted")

In that case, you might try (on the remote server) a git fsck --full 21f3981 ; git repack (as suggested by J-16 SDiZ)

You will find similar error message in this old (2008) thread, with Linus Torvald suggesting, for an error message like "unable to unpack .. header":

The "unable to unpack .. header" problem would at a guess be a totally corrupted loose object. You should have a file named

.git/objects/b2/8b949a1a3c8eb37ca6eefd024508fa8b253429

and it sounds like that file is corrupt.
So far, apart from a CRLF conversion bug (that you wouldn't have triggered on OS X anyway), I think every single time we've seen that, it's been a real disk or memory corruption.

Community
  • 1
  • 1
VonC
  • 1,262,500
  • 529
  • 4,410
  • 5,250
  • before that line it gets 62% through the compression, same percentage when I run `git gc` on the remote. – jondavidjohn Feb 16 '12 at 14:49
  • @jondavidjohn: so, no other error message at any line *before* the "died with error" message? – VonC Feb 16 '12 at 15:10
  • right, it gets about 62% through compression, stops, then outputs the above message. Also, as the question points out when I run `git fsck` i get no output, indicated that git has no knowledge of corrupted files or problems. correct? – jondavidjohn Feb 16 '12 at 15:56
  • @jondavidjohn: yes, correct. Just to be extra-sure, you did run the `fsck` on the remote server? Can you try on the remote a `git fsck --full`? – VonC Feb 16 '12 at 15:59
  • sure, yes that was on the remote server, I'll try the --full option. – jondavidjohn Feb 16 '12 at 16:20