Re: [git failure] failure pulling latest Linus tree

Re: [git failure] failure pulling latest Linus tree

[Petr Baudis]

Dear diary, on Wed, Oct 25, 2006 at 05:43:44PM CEST, I got a letter
where "H. Peter Anvin" <hpa@zytor.com> said that...
> HOWEVER, git 1.4.3 seems to have been bad chicken.  When we ran it we 
> got a neverending stream of segfaults in the logs.

Could you provide some details, please? Ideally a backtrace or
something? Also, it might be a good idea to use 1.4.3.2 since it fixes a
problem with compatibility of git-daemon and older git clients.

Thanks,

-- 
				Petr "Pasky" Baudis
Stuff: http://pasky.or.cz/
#!/bin/perl -sp0777i<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<j]dsj
$/=unpack('H*',$_);$_=`echo 16dio\U$k"SK$/SM$n\EsN0p[lN*1

Re: [git failure] failure pulling latest Linus tree

[H. Peter Anvin]

Petr Baudis wrote:
> Dear diary, on Wed, Oct 25, 2006 at 05:43:44PM CEST, I got a letter
> where "H. Peter Anvin" <hpa@zytor.com> said that...
>> HOWEVER, git 1.4.3 seems to have been bad chicken.  When we ran it we 
>> got a neverending stream of segfaults in the logs.
> 
> Could you provide some details, please? Ideally a backtrace or
> something? Also, it might be a good idea to use 1.4.3.2 since it fixes a
> problem with compatibility of git-daemon and older git clients.
> 

I'm afraid I can't anymore since I've flushed those binaries.  1.4.3.2 
works fine, it seems, but we hadn't gotten that update until yesterday.

Re: [git failure] failure pulling latest Linus tree

[Junio C Hamano]

"H. Peter Anvin" <hpa@zytor.com> writes:

> For some reason which we haven't been able to track down yet, the
> recent load imposed by FC6 caused zeus1's load to skyrocket, but not
> zeus2's... it's largely a mystery.

Would kernel.org prefer RPM cut on a FC6 box now?

> HOWEVER, git 1.4.3 seems to have been bad chicken.  When we ran it we
> got a neverending stream of segfaults in the logs.

If that is git-daemon dying when talking to older clients, that
has been fixed in 1.4.3.2 (it's virtual hosting support had an
off-by-one wrong check to tell older clients from newer one).

Sorry about that -- we heard about the incompatibility this
Monday.

Re: [git failure] failure pulling latest Linus tree

[H. Peter Anvin]

Junio C Hamano wrote:
> "H. Peter Anvin" <hpa@zytor.com> writes:
> 
>> For some reason which we haven't been able to track down yet, the
>> recent load imposed by FC6 caused zeus1's load to skyrocket, but not
>> zeus2's... it's largely a mystery.
> 
> Would kernel.org prefer RPM cut on a FC6 box now?
> 

No; since we have now (finally) archieved "OS parity" we're going to 
keep hera, zeus1, and zeus2 in sync going forward.

>> HOWEVER, git 1.4.3 seems to have been bad chicken.  When we ran it we
>> got a neverending stream of segfaults in the logs.
> 
> If that is git-daemon dying when talking to older clients, that
> has been fixed in 1.4.3.2 (it's virtual hosting support had an
> off-by-one wrong check to tell older clients from newer one).
> 
> Sorry about that -- we heard about the incompatibility this
> Monday.

This stuff happens... it's dealt with now.

	-hpa