threads / discuss / 1910

'bad file' error updating from Linus

Subject: 'bad file' error updating from Linus

## tl;dr

6 messages between Sep 22, 2005 and Sep 23, 2005.

replies: 5people: 4as markdown or json

walt· Sep 22, 2005, 11:40 UTC · lore

cg-update this morning is giving me this error:

receiving file list ... done v2.6.14-rc2

wrote 134 bytes read 1116 bytes 833.33 bytes/sec total size is 738 speedup is 0.59 Tree change: 044a500e46742d39d22f1781cfb64ba93b463e39:e98c07b9eabedfb1d7ebef61af475bdebe1dd64d error: unable to find e98c07b9eabedfb1d7ebef61af475bdebe1dd64d fatal: git-cat-file e98c07b9eabedfb1d7ebef61af475bdebe1dd64d: bad file error: unable to find e98c07b9eabedfb1d7ebef61af475bdebe1dd64d fatal: git-cat-file e98c07b9eabedfb1d7ebef61af475bdebe1dd64d: bad file Invalid tree id: e98c07b9eabedfb1d7ebef61af475bdebe1dd64d

Is this a kernel.org rsync problem, perhaps?

(I'm running the latest releases of cogito and git from kernel.org.)

porpen@gmail.com· Sep 22, 2005, 12:42 UTC · re: walt · lore

Re: 'bad file' error updating from Linus

You aren't alone. I had a look at http://www.kernel.org/git/ only to find that the 2.6 kernel has been taken off the list... /me thinks Linus is pushing an update as I type.

On a whim, I'll suggest we relax and go get some breakfast. ;-) and try again a little later...

Cheers! -P

On 9/22/05, walt <wa1ter@myrealbox.com> wrote:

Show 9 quoted lines

> cg-update this morning is giving me this error:
>
> receiving file list ... done
> v2.6.14-rc2
>
> wrote 134 bytes  read 1116 bytes  833.33 bytes/sec
> total size is 738  speedup is 0.59
> Tree change: 044a500e46742d39d22f1781cfb64ba93b463e39:e98c07b9eabedfb1d7ebef61af475bdebe1dd64d
> error: unable to find e98c07b9eabedfb1d7ebef61af475bdebe1dd64d

walt· Sep 22, 2005, 21:44 UTC · re: porpen@gmail.com · lore

Re: 'bad file' error updating from Linus

porpen@gmail.com wrote:

> You aren't alone.  I had a look at http://www.kernel.org/git/  only to
> find that the 2.6 kernel has been taken off the list...  /me thinks
> Linus is pushing an update as I type...

I think it was a disk failure, but no matter -- it's fixed now.

For the benefit of future googlers with the same 'bad file' error:

After I repeated the cg-update (successfully) I was still left with the 'bad file' error. I tried 'cg-restore' and the error went away. Very nifty!

Linus Torvalds· Sep 22, 2005, 23:32 UTC · re: walt · lore

Re: 'bad file' error updating from Linus

On Thu, 22 Sep 2005, walt wrote:

> 
> I think it was a disk failure, but no matter -- it's fixed now.

Nope, it's not.

It just depends on which IP address you happen to get when you look at www.kernel.org.

www.kernel.org resolves to two addresses:

	Name:   zeus-pub.kernel.org
	Address: 204.152.191.5
	Name:   zeus-pub.kernel.org
	Address: 204.152.191.37

and the .5 address is working, but the .37 address has the broken mirror.

I thought hpa took it out of service while it was re-syncing, but apparently not.

		Linus

H. Peter Anvin· Sep 23, 2005, 19:43 UTC · re: Linus Torvalds · lore

Re: 'bad file' error updating from Linus

Linus Torvalds wrote:

Show 18 quoted lines

> 
> Nope, it's not.
> 
> It just depends on which IP address you happen to get when you look at 
> www.kernel.org.
> 
> www.kernel.org resolves to two addresses:
> 
> 	Name:   zeus-pub.kernel.org
> 	Address: 204.152.191.5
> 	Name:   zeus-pub.kernel.org
> 	Address: 204.152.191.37
> 
> and the .5 address is working, but the .37 address has the broken mirror.
> 
> I thought hpa took it out of service while it was re-syncing, but 
> apparently not. 
>

Well, it had resynched, but the time and date stamp on the refs/heads/master file were correct - just the data wrong - so rsync didn't flag the file. I forced rsync --checksum and it's correct now.

	-hpa

walt· Sep 23, 2005, 00:56 UTC · re: walt · lore

Re: 'bad file' error updating from Linus

walt wrote: [...]

> After I repeated the cg-update (successfully) I was still left with
> the 'bad file' error.  I tried 'cg-restore' and the error went away.

Oops -- correction!

After encountering the same problem on a different machine, I find that 'cg-reset' is the one which fixed the error, not 'cg-restore'.

I believe that I tried both commands on the first machine, so I was uncertain which one actually did the fix.